The effect of a 2 week ketogenic diet, versus a carbohydrate‐based diet, on cognitive performance, mood and subjective sleepiness during 36 h of extended wakefulness in military personnel: An exploratory study

Extended wakefulness, or sleep deprivation, impairs cognitive performance and brain glucose metabolism. A ketogenic diet (KD) provides an alternative fuel source, ketone bodies, that could elicit a metabolic benefit during sleep deprivation. A randomised, cross‐over trial was conducted with seven male military personnel. Participants ingested an iso‐energetic ketogenic diet or carbohydrate‐based diet for 14 days, immediately followed by 36 h of extended wakefulness and separated by a 12 day washout. Cognitive performance, mood, subjective sleepiness, capillary blood glucose, and D‐β‐hydroxybutyrate concentrations were measured every 2 h during extended wakefulness. Linear mixed models were used to analyse data. D‐β‐hydroxybutyrate was higher (p < 0.001) and glucose was lower (p < 0.01) on the KD compared with the carbohydrate‐based diet. The KD improved psychomotor vigilance task performance (number of lapses, mean reciprocal response time, mean fastest 10% response time (RT), and mean slowest 10% RT; all p < 0.05), running memory continuous performance test performance (RT and number of correct responses per minute; both p < 0.01), and vigour, fatigue, and sleepiness (all, p ≤ 0.001) compared with the carbohydrate‐based diet. In conclusion, a KD demonstrated beneficial effects on cognitive performance, mood, and sleepiness during 36 h of extended wakefulness compared with a carbohydrate‐based diet.


Summary
Extended wakefulness, or sleep deprivation, impairs cognitive performance and brain glucose metabolism. A ketogenic diet (KD) provides an alternative fuel source, ketone bodies, that could elicit a metabolic benefit during sleep deprivation. A randomised, cross-over trial was conducted with seven male military personnel. Participants ingested an iso-energetic ketogenic diet or carbohydrate-based diet for 14 days, immediately followed by 36 h of extended wakefulness and separated by a 12 day washout. Cognitive performance, mood, subjective sleepiness, capillary blood glucose, and D-β-hydroxybutyrate concentrations were measured every 2 h during extended wakefulness. Linear mixed models were used to analyse data. D-β-hydroxybutyrate was higher (p < 0.001) and glucose was lower (p < 0.01) on the KD compared with the carbohydratebased diet. The KD improved psychomotor vigilance task performance (number of lapses, mean reciprocal response time, mean fastest 10% response time (RT), and mean slowest 10% RT; all p < 0.05), running memory continuous performance test performance (RT and number of correct responses per minute; both p < 0.01), and  (Caldwell et al., 2019). Cerebral carbohydrate (CHO) metabolism appears to decline during SD (Thomas et al., 2000), which coincides with reduced cognitive performance (Wu et al., 1991). As CHO is the primary energy source for the brain (Mergenthaler et al., 2013), the supply of alternative oxidisable energetic substrates, such as ketone bodies (KBs), could ameliorate cognitive impairments following SD.
This can increase the contribution of KBs to brain energy production and reduce the contribution of glucose (i.e., CHO) (Courchesne-Loyer et al., 2017). We recently demonstrated no effect of a KD on cognitive performance in military personnel after 1 and 2 weeks of adherence to a KD in a non-SD state (Shaw et al., 2022), which corroborated earlier research demonstrating that cognitive performance was unaltered after $29 days of a KD (Iacovides et al., 2019). Therefore, the cognitive benefit of a KD is potentially more likely to be observed following SD when the brain CHO metabolism is impaired and this remains to be investigated.
The aim of our study was to examine the effect of a 2-week KD, versus a CHO-based diet, on cognitive performance, mood, and subjective sleepiness during 36 h of extended wakefulness in military personnel. We hypothesised that SD-related impairments in cognitive performance, mood, and sleepiness would be mitigated by the KD compared with the CHO-based diet.

| METHODS
This randomised, controlled, cross-over trial was conducted with male military personnel. Written consent was obtained prior to commencement and participation was voluntary. Ethical approval was provided by the New Zealand Defence Force and Massey University Ethics Committees (SOA 20/47).
Participants ingested their habitual diet for 7 days during a baseline phase (data not reported here), then in the 14-day dietary adaptation phase the participants were randomised to either a KD (<5% energy intake [EI; <40 gÁday À1 ] from CHO, 15%-20% energy intake from protein and > 75% energy intake from fat) or CHO-based diet (>45% energy intake from CHO, 15%-20% energy intake from protein and < 40% energy intake from fat). Dietary compliance during dietary adaptation was measured using weight, image assisted diet records on three non-consecutive days each week, and verified using daily measures of waking, fasted capillary whole-blood D-βHB concentration (Shaw et al., 2022). Sleep was monitored using actigraphy Cognitive performance was assessed with the 10 minute psychomotor vigilance task (Dinges & Powell, 1985), and analysed for the
A detailed summary of dietary intake for each dietary condition during the adaptation phase is reported in Supporting Information, Table S3. Briefly, participants adhered to all dietary requirements and there were no differences between dietary conditions for EI; however, the KD was slightly ($21 g) higher in protein (p = 0.049). Blood D-βHB ± 1.1. The mean sleep duration for the 7 days prior to the extended wakefulness period did not differ between the KD (6.9 ± 0.9 hÁday À1 ) and CHO-based diet (6.7 ± 1.1 hÁday À1 ) (p = 0.20). On the night preceding the extended wakefulness period, participants slept for 6.8 Values are presented as mean ± standard deviation.
F I G U R E 2 Mean reciprocal response time (responses s À1 ) during 36 hours of extended wakefulness for the carbohydrate and ketogenic diet interventions. Values are presented as mean ± standard deviation.
during the dietary adaptation phase, as reported previously (Shaw et al., 2022). In support of our hypothesis, the KD attenuated SDrelated impairments for all PVT and RMCPT variables, mood, and subjective sleepiness during the 36 h. This suggests that increases in the availability of KBs and underlying metabolic adaptations to a shortterm KD may provide a metabolic benefit to the brain during SD.
In the present study, the interaction between increasing homeostatic sleep pressure and circadian processes elicited expected cognitive and mood impairments from which to assess possible effects of diet. Differences between dietary conditions were observed across the 36 h of extended wakefulness but appeared greatest in period 4 (0130-0630; that is, the circadian low). This was similar to a previous study comparing a 1 week, non-ketogenic (40% energy intake as CHO) high-fat diet with a high-CHO diet (65% energy intake from CHO) diet over 24 h of SD (Lowden et al., 2004), which found that irresistible sleepiness increased to a greater extent for the high-CHO diet during 1200-1600 and 2400-0400. Together, these observations suggest that the effects of a high fat, or KD, may be more beneficial during circadian lows.
We speculated that the metabolic effects of the KD would mitigate cognitive impairments that stemmed from reductions in brain glucose metabolism caused by SD. In the present study, adaptation to the KD lowered blood glucose concentration, which elicited hyperketonaemia via increased ketogenesis. Whilst we did not measure differences in brain metabolism, the 2-week adaptation period prior to

CONFLICT OF INTEREST
The authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.
F I G U R E 4 Vigour during 36 hours of extended wakefulness for the carbohydrate and ketogenic diet interventions. Values are presented as mean ± standard deviation.