Efficacy of a Web-Based Executive Function Training Program to Induce Healthier Food Choices and Increase Valuation...

Executive function (EF) is a set of explicit (top-down) cognitive abilities theorized to assist in conscious control of eating behavior. However, it is unclear whether EF training can improve valuation and consumption of nutrient-dense foods while concurrently altering attitudes toward and perceptions of nutrient-poor foods and their consumption.

The primary objective is to determine the efficacy of a web-based EF training program designed to concurrently improve attitudes toward and perceptions and consumption of nutrient-dense foods (fruits and vegetables) while worsening those of nutrient-poor foods (ie, candy and fried snacks).

Participants include a nationally representative US sample (N=500) of men and women aged 18 to 70 years with a BMI between 18.5 and 38 kg/m². Potential participants are recruited through an open link to a prescreening questionnaire. Qualified participants are sent an electronic version of the informed consent form to sign. After consenting, they are invited to create an account on our semiautomated web-based platform and randomized to EF training via a suite of 4 web-based games that promote responding to fruits and vegetables, inhibit responses to nutrient-poor foods, and train attention toward the former and away from the latter or sham training that involves images of flowers and songbirds (control). Training occurs at least twice per week. EF is tested before and after 8 weeks of training. Dietary intake is measured before and after the 8-week intervention. Group and sex differences in frequency of participants violating inhibitory signals will be assessed using log-linear models. Linear mixed-effects models will test whether web-based EF training improves EF as measured via response inhibition and attention during computer-based gameplay. The independent and interactive associations between inhibitory and attentional learning and consumption of nutrient-poor foods will be tested using mixed model regression. If random assignment results in baseline imbalances across conditions on any variables, they will be used as covariates in the mixed models. We will use both an intention-to-treat analysis using (restricted) maximum likelihood estimation and a completer analysis to understand the results from those participants who completed the training at least twice per week.

As of June 1, 2025, overall, 899 individuals have applied to participate in the study. Of 899 applicants, 763 (84.9%) qualified and 630 (70.1%) consented. There have been 386 withdrawals due to noncompliance with study requirements (n=342, 88.6%) or identification as a bot (n=35, 9.1%).

The association between improved EF and healthy eating behaviors may be key to improving diet quality. If effective, this web-based EF training program will provide a platform that can be made widely available to aid individuals in food-related decision-making. The platform could be modified to target other food choices and other health behaviors, such as reducing sedentariness and increasing physical activity.

ClinicalTrials.gov NCT05938894; https://clinicaltrials.gov/study/NCT05938894.

DERR1-10.2196/66394.