Previous studies have suggested beneficial effects of low glycemic index and low GL weight loss diets on mood and cognition [13
] and [30
]. However, studies to date have all been short-term, and there is no information on longer-term effects in individuals actually losing weight. The results of this unique investigation with provided food (and therefore highly controlled dietary intake) indicate no differential effects of HG and LG diets on cognitive performance during weight loss, and no adverse effects of weight loss generally for the parameters assessed. However, whereas some previous long-term studies have suggested negative effects of very low carbohydrate (and therefore low GL) ketogenic diets on mood, or positive effects of HG diets, the opposite trend was observed in this study. Specifically, we observed that volunteers randomized to the LG diet had no change in the depression subscale of POMS during weight loss, whereas volunteers randomized to the HG diet experienced a negative change. This study used a relatively small population and therefore further studies are needed, but the results are consistent with the suggestion that moderately lower carbohydrate and low GL eating patterns may be protective against negative mood change during weight loss whereas negative changes occur if a conventional HG diet is consumed. In addition, the results lend support to accumulating evidence of broadly beneficial health effects of low GL diets compared to HG diets in weight management [43
] and [45
Because this investigation of mood and cognition was part of a larger randomized controlled trial investigating the metabolic effects of CR, subjects were provided with all food, caloric beverages and a multivitamin supplement for 6 months. Thus, the effects of the different diets were studied under exceptionally controlled dietary conditions and are likely to be due to randomizations and consequent macronutrient intakes, rather than to potential confounding factors such as inadequate micronutrient intakes [46
]. It should be noted that, although data from two levels of CR (10% and 30%) were combined in the analyses with CR level used as a covariate, there was no significant effect of CR level on weight loss because 10%CR subjects did not typically consume all provided food.
Concerning our observations of effects of LG versus HG diets on mood, we observed these effects when measuring the acute rather than long-term effects of food on mood, out of concern that the long-term instruments may provide less sensitive indicators of mood change within the normal range. A significant diet by time interaction (p=0.04) for the depression subscale of the POMS, with depression increasing on the higher carbohydrate HG diet over time compared to no change on the lower carbohydrate LG diet. The magnitude of the changes did not result in any clinical levels of depression, and the functional effects of changes of the magnitude observed is not known and requires further study. However, to our knowledge, these results are the first to highlight a potentially beneficial effect of an LG diet on sub-clinical depression during weight loss.
Concerning underlying mechanisms influencing depression symptomology when consuming HG or LG diets, we speculated that hunger might be a factor in the results obtained. Previous studies by our group have shown that consumption of HG meals provokes adverse hormonal changes and alterations in the availability of metabolic fuels such as glucose and free fatty acids that exacerbate hunger [47
] and [45
]. In addition, increased fluctuations in blood glucose (which occur with consumption of HG diets) are associated with negative effects on mood in some [19
] and [49
] though not all [50
] studies. These observations, combined with reports that feelings of vigor are negatively related to hunger sensations [51
], and that consumption of extreme hunger-promoting diets is associated with negative mood [52
], made increased hunger when consuming an HG diet a potential candidate for the results obtained in this study. However, there was no effect of the HG versus LG diets on hunger (assessed with the Eating Inventory). Furthermore, controlling for hunger in models assessing the effect of diet in changes in mood parameters over time actually increased the significance of the diet by time interaction for depression (from p=0.04 to p=0.009). The negative mood effect of the HG diet in this study was thus apparently not due to increased hunger, but given the small size of our study further investigations are needed.
Our results are also relevant to the carbohydrate-depression hypothesis [53
] and [55
], which postulates that high carbohydrate diets improve mood through increased delivery of tryptophan (a precursor of the mood-modulating neurotransmitter serotonin) to the brain. Support for the hypothesis has been provided by self-reports of higher carbohydrate intake associated with reduced levels of depressive symptoms [20
] and [21
], but conversely meals containing as little as 4% protein appear to counteract any carbohydrate-induced rise in the plasma ratio of tryptophan to other large neutral amino acids that controls uptake into the brain [56
], and high-carbohydrate and high-protein breakfasts can cause substantial differences in plasma tryptophan ratios [57
]. However, if valid, the substantially higher carbohydrate HG diet tested in this study would be expected to yield more positive mood outcomes than the LG diet. In contrast, we saw the opposite effect, with our higher-carbohydrate HG diet resulting in a worsening of mood over time compared to the LG diet.
In contrast to the significant association between diet randomization and changes in the assessment of depression by POMS we observed no differences between dietary groups in changes in cognitive variables. In order to effectively investigate cognition, we selected a battery of cognitive tests that assessed a broad range of cognitive functions including reaction time, vigilance, learning, working memory and reasoning. Several previous studies have reported that weight loss is associated with poorer sustained attention, diminished immediate recall and longer simple reaction times [58
] and [59
], and suggested that the cognitive requirements of calorie counting and increased preoccupation with body weight may create measurable cognitive deficits in other areas [60
]. Despite such potentially negative suggested effects of weight loss on cognitive function, our results support other previous work suggesting no adverse effects of moderate CR on cognition [8
] and [12
]. These results were obtained in a subgroup of the whole study population, and require confirmation in a larger study. Nevertheless, the apparent negative effects of weight loss on cognitive function in other studies may be less reliable, due to confounding factors such as non compliance with dietary recommendations and possibly suboptimal micronutrient intakes [61
] and [63
In conclusion, we observed no effect of consuming a provided HG or LG weight-loss diet for 6 months on changes in cognitive function. However, randomization to the HG diet was associated with a relatively negative change in subclinical depression symptomology over time compared to randomization to the LG diet. Given that adverse changes in mood may negatively impact a long-term commitment to healthy eating habits and weight control, our findings suggest a potentially important psychological benefit of consuming LG diets for weight loss, and further studies in this area are warrented.