Automatic food detection in egocentric images using artificial intelligence technology.

OBJECTIVE: To develop an artificial intelligence (AI)-based algorithm which can automatically detect food items from images acquired by an egocentric wearable camera for dietary assessment. DESIGN: To study human diet and lifestyle, large sets of egocentric images were acquired using a wearable device, called eButton, from free-living individuals. Three thousand nine hundred images containing real-world activities, which formed eButton data set 1, were manually selected from thirty subjects. eButton data set 2 contained 29 515 images acquired from a research participant in a week-long unrestricted recording. They included both food- and non-food-related real-life activities, such as dining at both home and restaurants, cooking, shopping, gardening, housekeeping chores, taking classes, gym exercise, etc. All images in these data sets were classified as food/non-food images based on their tags generated by a convolutional neural network. RESULTS: A cross data-set test was conducted on eButton data set 1. The overall accuracy of food detection was 91·5 and 86·4 %, respectively, when one-half of data set 1 was used for training and the other half for testing. For eButton data set 2, 74·0 % sensitivity and 87·0 % specificity were obtained if both 'food' and 'drink' were considered as food images. Alternatively, if only 'food' items were considered, the sensitivity and specificity reached 85·0 and 85·8 %, respectively. CONCLUSIONS: The AI technology can automatically detect foods from low-quality, wearable camera-acquired real-world egocentric images with reasonable accuracy, reducing both the burden of data processing and privacy concerns.

The SMARTER pilot study: Testing feasibility of real-time feedback for dietary self-monitoring.

Self-monitoring (SM) of food intake is central to weight loss treatment. Technology makes it possible to reinforce this behavior change strategy by providing real-time feedback (FB) tailored to the diary entry. To test the feasibility of providing 1-4 daily FB messages tailored to dietary recordings via a smartphone, we conducted a 12-week pilot randomized clinical trial in Pittsburgh, PA in US in 2015. We compared 3 groups: SM using the Lose It! smartphone app (Group 1); SM + FB (Group 2); and SM + FB + attending three in-person group sessions (Group 3). The sample (N = 39) was mostly white and female with a mean body mass index of 33.76 kg/m2. Adherence to dietary SM was recorded daily, weight was assessed at baseline and 12 weeks. The mean percentage of days adherent to dietary SM was similar among Groups 1, 2, and 3 (p = 0.66) at 53.50% vs. 55.86% vs. 65.33%, respectively. At 12 weeks, all groups had a significant percent weight loss (p < 0.05), with no differences among groups (- 2.85% vs. - 3.14% vs. - 3.37%) (p = 0.95); 26% of the participants lost ≥ 5% of their baseline weight. Mean retention was 74% with no differences among groups (p = 0.37). All groups adhered to SM at levels comparable to or better than other weight loss studies and lost acceptable amounts of weight, with minimal intervention contact over 12 weeks. These preliminary findings suggest this 3-group approach testing SM alone vs. SM with real-time FB messages alone or supplemented with limited in-person group sessions warrants further testing in a larger, more diverse sample and for a longer intervention period.

Ecological Momentary Assessment in Behavioral Research: Addressing Technological and Human Participant Challenges.

BACKGROUND: Ecological momentary assessment (EMA) assesses individuals' current experiences, behaviors, and moods as they occur in real time and in their natural environment. EMA studies, particularly those of longer duration, are complex and require an infrastructure to support the data flow and monitoring of EMA completion. OBJECTIVE: Our objective is to provide a practical guide to developing and implementing an EMA study, with a focus on the methods and logistics of conducting such a study. METHODS: The EMPOWER study was a 12-month study that used EMA to examine the triggers of lapses and relapse following intentional weight loss. We report on several studies that informed the implementation of the EMPOWER study: (1) a series of pilot studies, (2) the EMPOWER study's infrastructure, (3) training of study participants in use of smartphones and the EMA protocol and, (4) strategies used to enhance adherence to completing EMA surveys. RESULTS: The study enrolled 151 adults and had 87.4% (132/151) retention rate at 12 months. Our learning experiences in the development of the infrastructure to support EMA assessments for the 12-month study spanned several topic areas. Included were the optimal frequency of EMA prompts to maximize data collection without overburdening participants; the timing and scheduling of EMA prompts; technological lessons to support a longitudinal study, such as proper communication between the Android smartphone, the Web server, and the database server; and use of a phone that provided access to the system's functionality for EMA data collection to avoid loss of data and minimize the impact of loss of network connectivity. These were especially important in a 1-year study with participants who might travel. It also protected the data collection from any server-side failure. Regular monitoring of participants' response to EMA prompts was critical, so we built in incentives to enhance completion of EMA surveys. During the first 6 months of the 12-month study interval, adherence to completing EMA surveys was high, with 88.3% (66,978/75,888) completion of random assessments and around 90% (23,411/25,929 and 23,343/26,010) completion of time-contingent assessments, despite the duration of EMA data collection and challenges with implementation. CONCLUSIONS: This work informed us of the necessary preliminary steps to plan and prepare a longitudinal study using smartphone technology and the critical elements to ensure participant engagement in the potentially burdensome protocol, which spanned 12 months. While this was a technology-supported and -programmed study, it required close oversight to ensure all elements were functioning correctly, particularly once human participants became involved.

Trajectories of Weight Change and Predictors Over 18-Month Weight Loss Treatment.

BACKGROUND: Obesity research has typically focused on weight change patterns using the whole sample in randomized clinical trials (RCTs), ignoring subsets of individuals with varying weight change trajectories (e.g., continuing to lose, or maintaining weight). The purpose was to explore possible trajectories of weight change and their associated predictors. METHODS: We conducted a secondary analysis of data from two RCTs using standard behavioral treatment for weight loss. Group-based trajectory modeling was used to identify distinct classes of percent weight change trajectories over 18 months. RESULTS: The sample (N = 338) was primarily female (85.2%), White (73.7 %), 45.7 ± 9.0 years old, with 15.6 ± 2.8 years of education. Three trajectory groups were identified: good responders (>15% weight loss), fair responders (5%-10% weight loss), and poor responders (<5% weight loss). The good responders had a significantly larger decrease in perceived Barriers to Healthy Eating subscale scores than the fair and poor responders (p < .01). Compared to the poor responders, there was a significant decrease in fat gram intake in the good responders (p = .01). CONCLUSIONS: Good responders differed from poor responders in decreasing their perceived barriers to healthy eating (e.g., managing emotions, social support, and daily mechanics of adopting a healthy diet) and reducing fat intake. Good responders differed from fair responders in perceived barriers to healthy eating. CLINICAL RELEVANCE: Clinicians need to focus on how we can assist those who are being unsuccessful in adopting some of the behaviors observed among those who have experienced successful weight loss and maintainers.

Socio-demographic, anthropometric, and psychosocial predictors of attrition across behavioral weight-loss trials.

Preventing attrition is a major concern in behavioral weight loss intervention studies. The purpose of this analysis was to identify baseline and six-month predictors associated with participant attrition across three independent clinical trials of behavioral weight loss interventions (PREFER, SELF, and SMART) that were conducted over 10 years. Baseline measures included body mass index, Barriers to Healthy Eating, Beck Depression Inventory-II (BDI), Hunger Satiety Scale (HSS), Binge Eating Scale (BES), Medical Outcome Study Short Form (MOS SF-36 v2) and Weight Efficacy Lifestyle Questionnaire (WEL). We also examined early weight loss and attendance at group sessions during the first 6 months. Attrition was recorded at the end of the trials. Participants included 504 overweight and obese adults seeking weight loss treatment. The sample was 84.92% female and 73.61% white, with a mean (± SD) age of 47.35 ± 9.75 years. After controlling for the specific trial, for every one unit increase in BMI, the odds of attrition increased by 11%. For every year increase in education, the odds of attrition decreased by 10%. Additional predictors of attrition included previous attempts to lose 50-79 lbs, age, not possessing health insurance, and BES, BDI, and HSS scores. At 6 months, the odds of attrition increased by 10% with reduced group session attendance. There was also an interaction between percent weight change and trial (p<.001). Multivariate analysis of the three trials showed education, age, BMI, and BES scores were independently associated with attrition (ps ≤ .01). These findings may inform the development of more robust strategies for reducing attrition.

The SELF trial: A self-efficacy-based behavioral intervention trial for weight loss maintenance.

OBJECTIVE: The SELF Trial examined the effect of adding individual self-efficacy (SE) enhancement sessions to standard behavioral weight loss treatment (SBT). METHODS: Participants were randomly assigned to SBT or SBT plus SE sessions (SBT+SE). Outcome measures were weight loss maintenance, quality of life, intervention adherence, and self-efficacy at 12 and 18 months. RESULTS: The sample (N = 130) was female (83.08%) with a mean (SD) body mass index of 33.15 (4.11) kg m(2) . There was a significant time effect for percent weight change (P = 0.002) yet no significant group or group-by-time effects. The weight loss for the SBT+SE group was 8.38% (7.48) at 12 months and 8.00% (7.87) at 18 months, with no significant difference between the two time points (P = 0.06). However, weight loss for the SBT group was 6.95% (6.67) at 12 months and 5.96% (7.35) at 18 months, which was significantly different between the two time points (P = 0.005), indicating that the SBT group had significant weight regain. CONCLUSIONS: Both groups achieved clinically significant weight loss. The group receiving an intervention targeting enhanced self-efficacy had greater weight loss maintenance whereas the SBT group demonstrated significant weight regain possibly related to the greater attention provided to the SBT+SE group.

Changes in adipose tissue depots and metabolic markers following a 1-year diet and exercise intervention in overweight and obese patients with type 2 diabetes.

OBJECTIVE: We aim to characterize the effects on total body fat and distribution of a 1-year intensive lifestyle intervention (ILI) for weight loss in overweight and obese adults with type 2 diabetes and to examine whether changes in adipose tissue (AT) depots were associated with changes in metabolic biomarkers. RESEARCH DESIGN AND METHODS: Participants were 54 females and 38 males (age 57.8 ± 6.7 years [mean ± SD]; BMI 31.7 ± 3.5 kg/m(2)) enrolled in the Look AHEAD (Action for Health in Diabetes) trial randomized to ILI or diabetes support and education (DSE) from whom baseline and 1-year MRI measures of total AT (TAT) and regional (arm, trunk, leg) AT, including subcutaneous AT (SAT), visceral AT (VAT), and intermuscular AT (IMAT), were acquired. We tested whether mean changes in ILI and DSE were equal and, within groups, whether changes were different from zero. Regression models tested whether changes in AT compartments were associated with metabolic variable changes. RESULTS: Body weight changed -0.52 ± 3.62 kg (P = 0.31) in DSE and -7.24 ± 5.40 kg (P < 0.0001) in ILI. Mean ILI changes were different from DSE (P < 0.001 for TAT, SAT, and IMAT and P < 0.01 for VAT in females). Within ILI, SAT and VAT decreased in males and females (P < 0.0001), but IMAT was unchanged (0.00 ± 0.54 kg; P = 0.99). In DSE, SAT and VAT did not change, but IMAT increased by 0.46 ± 0.55 kg (P < 0.001). Controlling for weight loss, reduction of specific AT depots was associated with improvement in metabolic biomarkers. CONCLUSIONS: Weight loss of 7-10% from an ILI over 1 year reduced SAT and VAT and prevented an increase in IMAT. Reductions in AT depots were associated with improvements in biomarkers.

Measurements of islet function and glucose metabolism with the dipeptidyl peptidase 4 inhibitor vildagliptin in patients with type 2 diabetes.

OBJECTIVE: Pharmacological inhibition with the dipeptidyl peptidase 4 (DPP-4) inhibitor vildagliptin prolongs the action of endogenously secreted incretin hormones leading to improved glycemic control in patients with type 2 diabetes mellitus (T2DM). We undertook a double-blinded, randomized-order, crossover study to examine the vildagliptin mechanisms of action on islet function and glucose utilization. RESEARCH DESIGN AND METHODS: Participants with T2DM (n = 16) who had a baseline hemoglobin A(1c) of 7.1 +/- 0.2% completed a crossover study with 6 wk of treatment with vildagliptin and 6 wk with placebo. At the completion of each arm, participants had a study of postprandial metabolism and a two-step glucose clamp performed at 20 and 80 mU/min x m(2) insulin infusions. RESULTS: Vildagliptin increased postprandial glucagon-like peptide-1 and glucose-dependent insulinotropic polypeptide by 3- and 2-fold, respectively, reduced fasting plasma glucose and postprandial plasma glucose by 1.3 +/- 0.3 mmol/liter and 1.6 +/- 0.3 mmol/liter (both P <0.01), and improved glucose responsiveness of insulin secretion by 50% (P < 0.01). Vildagliptin lowered postprandial glucagon by 16% (P <0.01). Examined by glucose clamp, insulin sensitivity and glucose clearance improved after vildagliptin (P < 0.01). CONCLUSIONS: Vildagliptin improves islet function in T2DM and improves glucose metabolism in peripheral tissues.

Orlistat: Current issues for patients with type 2 diabetes.

Orlistat, a pancreatic lipase inhibitor, was approved by the US Food and Drug Administration (FDA) in the spring of 1999 as an adjunct to lifestyle intervention for weight loss. This paper seeks to examine current issues regarding orlistat use in patients with type 2 diabetes. There are a number of trials that demonstrate the benefits of orlistat over placebo for reducing body weight and improving other health parameters. Of some interest are the preliminary explorations of interaction on cytokine levels, where a possible cardiovascular benefit is plausible. Implications of the FDA approval of over-the-counter use and the pharmaceutical development of another lipase inhibitor are also examined.

Effect of weight loss and nutritional intervention on arterial stiffness in type 2 diabetes.

OBJECTIVE: There is increased stiffness of the large central arteries in type 2 diabetic patients, and obesity is a risk factor. However, the effect of intentional weight loss on arterial stiffness is uncertain, and the purpose of the current study was to assess this effect. RESEARCH DESIGN AND METHODS: Arterial stiffness was assessed by measuring aortic pulse wave velocity (aPWV) at baseline and at completion of a 1-year weight loss intervention. Metabolic control of type 2 diabetes was also appraised. RESULTS: Mean weight loss at 1 year in 38 volunteers with type 2 diabetes was 7.8%. There were improvements in HbA1c, LDL cholesterol, homeostasis model assessment of insulin resistance, and inflammatory markers (plasminogen activator inhibitor-1, tumor necrosis factor-alpha, interleukin-6, and C-reactive protein). There was also a significant improvement in aPWV at completion of weight loss intervention, from 740 to 690 cm/s (P < 0.05). CONCLUSIONS: Moderate weight loss improves arterial stiffness in type 2 diabetes.

Effects of moderate weight loss and orlistat on insulin resistance, regional adiposity, and fatty acids in type 2 diabetes.

OBJECTIVE: Moderate weight loss is recommended for overweight and obese patients with type 2 diabetes, and conjunctive use of weight loss medication has been advocated. The current study examined weight loss-dependent and -independent effects of the intestinal lipase inhibitor orlistat at 6 months of treatment, using behavioral intervention (Int) combined with randomized, double-blinded, placebo (P)-controlled treatment with orlistat (O). RESEARCH DESIGN AND METHODS: Metabolic control, insulin sensitivity (IS), regional fat distribution, and fat content in liver and muscle were measured in 39 volunteers with type 2 diabetes in whom all antidiabetic medication was withdrawn 1 month preceding randomization. Weight loss was equivalent in the Int+O and Int+P groups, respectively (-10.3 +/- 1.3 vs. -8.9 +/- 1.1%), and there were identical decreases in visceral adipose tissue (VAT), fat mass (FM), thigh adiposity, and hepatic steatosis. RESULTS: Weight loss resulted in substantial improvement (P < 0.001) in HbA(1c) (-1.6 +/- 0.3 vs. -1.0 +/- 0.4%; NS between groups). IS improved significantly more with orlistat (Delta2.2 +/- 0.4 vs. Delta1.2 +/- 0.4 mg. min(-1). kg(-1) fat-free mass [FFM]; P < 0.05), and plasma free fatty acid (FFA) levels were strongly correlated with IS (r = 0.56; P < 0.001). Orlistat caused greater reductions in fasting plasma FFA (Delta-154 +/- 22 vs. Delta-51 +/- 33 micro mol/l; P < 0.05), insulin-suppressed FFA (Delta-119 +/- 23 vs. Delta-87 +/- 34 micro mol/l; P < 0.05), and fasting plasma glucose (FPG; -62 +/- 9 vs. -32 +/- 8 mg/dl; P = 0.02). Changes in HbA(1c) were correlated with DeltaIS (r = -0.41; P < 0.01) but not with weight loss per se. CONCLUSIONS: At equivalent weight loss, conjunctive use of orlistat resulted in greater improvement in FFA levels and IS.