Search

Background: Training of machine learning algorithms on dish images collected in other countries requires possible sources of systematic discrepancies, including country-specific food composition databases (FCDBs), to be tackled. The US Nutrition5k project provides for ~5000 dish images and related dish- and ingredient-level information on mass, energy, and macronutrients from the US FCDB. The aim of this study is to (1) identify challenges/solutions in linking the nutritional composition of Italian foods with food images from Nutrition5k and (2) assess potential differences in nutrient content estimated across the Italian and US FCDBs and their determinants., Methods: After food matching, expert data curation, and handling of missing values, dish-level ingredients from Nutrition5k were integrated with the Italian-FCDB-specific nutritional composition (86 components); dish-specific nutrient content was calculated by summing the corresponding ingredient-specific nutritional values. Measures of agreement/difference were calculated between Italian- and US-FCDB-specific content of energy and macronutrients. Potential determinants of identified differences were investigated with multiple robust regression models., Results: Dishes showed a median mass of 145 g and included three ingredients in median. Energy, proteins, fats, and carbohydrates showed moderate-to-strong agreement between Italian- and US-FCDB-specific content; carbohydrates showed the worst performance, with the Italian FCDB providing smaller median values (median raw difference between the Italian and US FCDBs: -2.10 g). Regression models on dishes suggested a role for mass, number of ingredients, and presence of recreated recipes, alone or jointly with differential use of raw/cooked ingredients across the two FCDBs., Conclusions: In the era of machine learning approaches for food image recognition, manual data curation in the alignment of FCDBs is worth the effort.

A non-invasive and non-pharmacological approach is evaluated for the proprioceptive and postural improvement of PD subjects. The authors evaluated the effectiveness of a class I medical device according to EU regulation 745/2017 designed to develop the mechanism of action based on the modulation of action potentials, which occurs in prevalent pathways of the afferent peripheral nervous system efferent in subjects with spasticity. The present observational study, structured in a double-blind randomized manner, therefore, had the main aim of evaluating the ability of the device to improve on the motor and proprioceptive function of PD patients. This study was based on the instrumented gait analysis performed according to the Timed Up and Go (TUG) test procedure, as well as using a fall risk assessment in accordance with the Berg Balance Scale (BBS) procedures. This study involved 25 participants in the active group (no placebo) and 25 in the non-active group (placebo), the latter to whom non-functional devices were applied, but in every respect identical to the functional devices applied to the 25 patients in the no placebo group. Data analysis was conducted using statistical methodologies for statistics, the statistical significance of the results for the observed samples and the interdependence between the measured variables. The study of the mechanism of action based on the remodulation of action potentials was preliminary conducted through numerical modeling of the Hodgkin-Huxley axon, modified by introducing the influence of the capacitive device applied in clinical tests into the validated model to target the dielectric properties of materials constituting the passive sensor. The use of the neuromodulation device promises observable improvements in motor function among PD patients, including increased limb mobility and greater postural stability.

Objective: The availability of nutrition applications (apps) has increased in recent years. The aim of this study was to assess the accuracy of nutrient intake calculations from some of the leading apps., Methods: We identified five apps according to some selection criteria: >4-star ratings, >1 million downloads, including a food composition database, and in Italian language. Apps were used for 2 wk each. Using a 3-d food diary, the nutritional values obtained from each app were compared to a reference method including the Food Composition Database for Epidemiologic Studies in Italy. Energy intake differences were calculated for single nutrient and 3-d food diary between single app and reference method after food-item matching. Bland-Altman plots were used to assess agreement of the methods., Results: Apps identified were FatSecret, Lifesum, MyFitnessPal, Yazio, and Melarossa. Apps tended to underestimate total energy intake compared with the reference method, from a minimum of -2 kcal for Lifesum, to a maximum of -5.4 kcal for Yazio (average per item). Apps tended to underestimate lipids, and to a lesser extent carbohydrate and fiber intake, except for Yazio and Lifesum, which overestimated the intake of protein. These discrepancies appear to be due to the use of no country-specific food composition databases and to user customization of the food list., Conclusions: The present findings suggest that the leading nutrition apps present critical issues in assessing the intake of energy and nutrients. Implementation of a framework for quality assessment is necessary to drive the design and development of higher-quality apps. Further research on efficacy and use of apps to monitor food intake is also warranted and some recommendations are provided., (Copyright © 2020 Elsevier Inc. All rights reserved.)