Short-term effects of carbohydrates differing in glycemic index (GI) consumed at lunch on children's cognitive function in a randomized crossover study

Children in all-day schools must sustain cognitive performance into the afternoon, and lunch composition may contribute to this. Glucose is the brain’s primary fuel and continuous supply is important, yet rapid, high postprandial glucose excursions could impair cognition, while slower, prolonged rises may benefit attention and memory. The glycemic index (GI) ranks carbohydrate-rich foods by their postprandial glycemic impact. Prior studies indicate that low-GI breakfasts can improve children’s cognition versus skipping breakfast or consuming high-GI breakfasts, though results are inconsistent across studies, potentially due to methodological differences and confounding. Evidence on lunch dietary GI and short-term cognitive effects in children has been lacking. Prior CogniDo studies found no negative short-term effects of having lunch on children’s cognition (contrasting adult postprandial fatigue), and suggested possible temporary benefits for working memory updating. This study aimed to test whether lunch dietary GI acutely affects cognitive performance in schoolchildren.

Previous research shows mixed effects of dietary GI on cognition in children. Studies reported better cognitive performance after low-GI breakfasts compared with skipping or high-GI breakfasts, yet at least one repeated-measures study found no breakfast GI effects, likely reflecting methodological variability and confounding. While adult studies have described postprandial fatigue after lunch, prior CogniDo interventions in children did not observe such declines and suggested small positive effects on certain domains. There had been no studies specifically addressing lunch dietary GI and short-term cognitive outcomes in children prior to this work.

Design: Randomized, single-blind, 2×2 crossover intervention integrated into the school routine over 10 weeks (April to mid-June 2016) at the all-day Comprehensive School Berger Feld, Gelsenkirchen, Germany. Participants: 5th and 6th graders (12 classes). Exclusions: metabolic disorders, epilepsy, special diet; post hoc exclusions for diagnosed learning disorder or insufficient German language skills. Randomization: Block randomization per class (blocks of 2–4) to sequence medium–high GI (m–hGI) or high–medium GI (h–mGI); computer-generated allocation. Two test days per participant scheduled one week apart (one class had 3 weeks due to cancellation). Interventions: Lunch consisted of rice with ground beef sauce; GI of candidate rice and pasta products was measured by SUGIRS (ISO 26642:2010). Two rice types with GI 62 (medium) and GI 86 (high) were selected (greater GI separation than pasta). In period 1, one sequence received hGI rice and the other mGI rice ad libitum; vice versa in period 2. Standardization and timing: Each test day started at 09:15 with a standardized breakfast (wholemeal bread, margarine, salami or Gouda, carrot sticks) ad libitum. At 12:25 (start of lunch break), participants received the assigned lunch. Intake assessment by weighing plates before and after. Estimated meal glycemic load (GL) for rice was calculated as GI × grams of rice carbohydrate per portion / 100; carbohydrates from the sauce were not included. Water available ad libitum. Cognitive assessment began at 13:15 (~45 min after lunch start). Participants were asked to refrain from foods and caloric beverages between breakfast and lunch; adherence monitored and a questionnaire on intake completed each intervention day. Cognitive outcomes: Computerized test battery (ALA Institute) administered in fixed order after explanation and practice: (1) Task switching: alternative Trail Making Task with three sections (numbers, letters, switch condition). Outcomes: total reaction time (RT) for numbers (items 2–26), total RT for letters (items 2–13), and switch costs (section 3 time minus section 1 time minus the difference between first 12 items of section 2 and section 1). Implausible negative switch costs excluded. (2) Working memory updating: two-back task with 106 stimuli (fruits/vegetables); stimuli 500 ms, ISI 2100 ms, max RT 1400 ms; 21 targets. Outcomes: ratio of missings (no response on target), ratio of false alarms (response on non-target), and mean RT on targets. (3) Tonic alertness: simple reaction task with 50 items; fixation cross followed by circle within 3300 ms (±20%); max RT 1500 ms. Outcomes: mean RT (ms), RT deviation (ms), omission errors (no response within 1500 ms), commission errors (response during fixation). Statistical analysis: SPSS v25; two-sided p < 0.05. Outcome variables treated as interval-scaled. Following CONSORT and Wellek & Blettner recommendations, sums of period 1 and 2 values were compared between sequences using unpaired t test or Wilcoxon rank-sum to assess carryover. If no carryover, treatment effects were estimated by comparing individual within-subject differences (day 1 − day 2) between sequences (hGI–mGI vs mGI–hGI). If carryover detected, only period 1 results were analyzed. Associations of GI with cognitive parameters were further analyzed using a linear mixed model adjusting for estimated meal GL and period; GI, GL, and period as fixed effects; subject as random effect.

- Participants: 193 consented; 4 excluded for learning disorder; 1 did not eat lunch on one day; intention-to-treat analysis included 188 students (5th–6th grade). - Intervention GI values: medium-GI rice GI = 62; high-GI rice GI = 86. - Primary crossover analysis: No significant effects of lunch dietary GI on most cognitive parameters in early afternoon across the two periods. - Carryover effects: Detected for two outcomes—two-back task reaction time and alertness task commission errors—necessitating analysis restricted to period 1 for these measures. - Period 1 effects (hGI vs mGI): • Two-back mean RT: faster in hGI group (p = 0.001). • Alertness task commission errors: fewer in hGI group (p = 0.04). - Period effects (both groups): switch costs improved in period 2; RT for visual search numbers slowed; in two-back and alertness tasks, RT and false alarm rates decreased while missings increased in period 2. - Meal consumption and GL: Amount consumed did not differ significantly between periods; estimated meal GL differed significantly between periods within sequences (Table 1 shows GL higher in period 2 for m–hGI sequence and lower in period 2 for h–mGI sequence). - Mixed model adjustment: Including estimated meal GL as covariate showed no significant associations of GI or GL with cognitive parameters. - Per-protocol sensitivity (adherent participants): Results consistent with primary analysis.

This school-based randomized crossover trial found no overall short-term effect of lunch dietary GI on cognitive performance in children when assessed approximately 45 minutes after lunch. Although prior research suggested potential cognitive benefits of lower postprandial glycemia (e.g., after low-GI breakfasts), this study did not observe consistent advantages of medium-GI versus high-GI rice at lunch. Carryover effects emerged for two measures, leading to reliance on first-period data for those outcomes; in that restricted analysis, high-GI rice was associated with faster working memory updating (two-back RT) and fewer commission errors in alertness, suggesting possible specific benefits that run counter to the expectation that lower GI would be advantageous. Period effects across multiple outcomes indicate practice or temporal influences independent of GI. Adjustment for estimated meal GL and period effects did not reveal significant associations of GI or GL with cognition, reinforcing the primary null findings. Overall, the results suggest that, in the tested timeframe and context, the GI of a rice-based lunch may not meaningfully alter early afternoon cognitive function in schoolchildren, though isolated parameters may be sensitive and warrant further investigation.

In this randomized crossover study in schoolchildren, lunch dietary GI (medium vs high) showed no consistent short-term effects on cognitive performance measured shortly after lunch. Exploratory findings of faster working memory updating and fewer commission errors following high-GI rice in first-period analyses require replication. Future research should confirm these specific effects, explore different postprandial time windows, meal compositions and macronutrient balances, and assess broader cognitive domains and longer-term academic outcomes.

- Carryover effects were present for two cognitive parameters (two-back RT and alertness commission errors), limiting the use of full crossover data and necessitating analysis restricted to period 1 for these outcomes. - Period effects (practice or timing) influenced several measures, complicating interpretation of treatment effects. - Estimated meal glycemic load differed between periods, and GL calculations excluded carbohydrates from the sauce, potentially reducing precision of exposure estimation. - The study was conducted in a single comprehensive school with assessments integrated into routine school days, which may limit generalizability. - One class had a longer interval between periods (3 weeks) due to scheduling, introducing minor variability in washout timing.