TABLE 1.

Summary of the research recommendations for studies examining the effect of breakfast on cognition¹

Limitation	Recommendations	Implications
Acute intervention studies
Sampling and sample selection
Overrepresentation of children	Research is required in adolescent samples.	Improves the generalizability of the findings to adolescents.
Small unrepresentative samples; lack of  power calculations	Research is required from sufficiently powered studies that use larger samples. Power calculations based on effect sizes demonstrated in previous research vs. effect size conventions are required.	Improves the generalizability of the findings. Improves confidence that null findings on CF are due to true lack of effect vs. lack of power.
Overrepresentation of children of a  mid- to high SES	Research is required in more at-risk populations, such as populations of a low SES and those with poorer cognitive ability.	Improves the generalizability of the findings. Increases the sensitivity of the intervention on CF.
Study location
Lack of field-based studies	More research is required from field-based studies, e.g., in schools.	Improves the ecological validity of the findings.
Breakfast manipulation
Lack of realistic breakfast manipulations	Research is required that uses ad libitum breakfast manipulations.	Improves the ecological validity of the findings. Benefits to CF may be more demonstrable with test meals that resemble habitual meals.
CF tests
Lack of studies that use sensitive CF tests	Research is required that uses tests with proven sensitivity to similar acute nutritional manipulations (e.g., serial sevens, free word recall, and cued word recall). Research required with more focused testing batteries that examine domains facilitated more reliably by breakfast consumption (attention, executive function, memory).	Improves confidence that null findings are due to true lack of effect vs. test insensitivity.
Chronic intervention studies (SBPs)
Study design
Poor-quality study design	Research is required from RCTs. Policy-makers responsible for SBPs should collaborate with researchers in the early design stages, before roll-out, to allow for a robust evaluation.	Improves the internal validity of the findings.
Poor or unreported effectiveness in  increasing breakfast consumption	Research is required that measures breakfast consumption and attendance at SBPs. Interventions should adopt an evidence-based theoretical framework to produce change in breakfast eating.	Improves confidence that the null findings are due to true lack of effect vs. lack of effectiveness in increasing breakfast consumption.
Isolation of acute vs. chronic effects
Temporal positioning of CF tasks  unreported or administered  postbreakfast	Effects should be measurable under fasting conditions after repeated consumption of the breakfast over time.	Allows for the isolation of acute effects from chronic effects.
Cross-sectional studies
Assessment and definition of HBC
Unvalidated brief dietary assessments	Research is required with the use of validated dietary measures that assess both frequency and composition of breakfast intake (e.g., food diary).	Improves the internal validity of the findings.
			Allows for data on both the frequency and composition of breakfast to be considered in the analysis.
Breakfast-eating occasion not defined	Research is required that adopts a standardized definition of breakfast: the breakfast-eating occasion should be defined for participants, or researchers should use a definition of breakfast post hoc (e.g., threshold amount of food and/or time of day).	Reduces inconsistencies between participants.
Confounding
Lack of control for confounders	Research-driven selection and accurate measurement of a range of confounders to include in the analysis are required.	Reduces residual and unmeasured confounding. Improves the internal validity of the findings.

¹CF, cognitive function; HBC, habitual breakfast consumption; RCT, randomized controlled trial; SBP, school breakfast program; SES, socioeconomic status.