Three consecutive weeks of nutritional ketosis has no effect on cognitive function, sleep, and mood compared with a high-carbohydrate, low-fat diet in healthy individuals: a randomized, crossover, controlled trial.
Level 2 - randomized trial
Individual randomized crossover controlled trial
PubMed 31098615 · doi:10.1093/ajcn/nqz073
What was done
Eleven healthy, normal-weight adults (mean age 30 ± 9 years) completed a randomized, controlled, crossover trial comparing two isocaloric diets followed for at least 3 weeks each, separated by a 1-week washout. The diets were a high-carbohydrate, low-fat (HCLF) diet (55% carbohydrate, 20% fat, 25% protein) and a ketogenic diet (KD) (15% carbohydrate, 60% fat, 25% protein). Nutritional ketosis was verified by blood β-hydroxybutyrate measurements. Cognitive performance across multiple domains (vigilance, visual learning/memory, working memory, executive function) was measured before and after each intervention using a validated computer-based test battery. Subjective mood and sleep were assessed throughout using validated rating scales.
What was found
Blood β-hydroxybutyrate confirmed ketosis during the KD phase (1.0 ± 0.5 mmol/L vs. baseline 0.2 ± 0.2 mmol/L, washout 0.2 ± 0.1 mmol/L, and HCLF 0.3 ± 0.2 mmol/L). Compared to the HCLF diet, 3 weeks of KD produced no significant differences in task speed or accuracy across tested cognitive domains: - Vigilance: speed P = 0.39 (Cohen's d = 0.26), accuracy P = 0.99 (d = 0.04) - Visual learning and memory: speed P = 0.99 (d = 0.04), accuracy P = 0.99 (d = 0.03) - Working memory: speed P = 0.62 (d = 0.26), accuracy P = 0.98 (d = 0.07) - Executive function: speed P = 0.60 (d = 0.31), accuracy P = 0.90 (d = 0.19) Mood, sleep quality, and morning vigilance also did not differ significantly between diets (all P > 0.05).
Why it matters
These findings indicate that short-term nutritional ketosis neither improves nor impairs cognitive performance, sleep, or mood in healthy, normal-weight individuals relative to an isocaloric high-carbohydrate diet.
Limits
The sample size was very small (n = 11), leaving the trial underpowered to detect subtle effects. The intervention was brief (3 weeks per diet), sleep was evaluated only via subjective self-reports rather than objective polysomnography, and the results from young, healthy, normal-weight participants cannot be generalized to older individuals or clinical populations.