3.1.1. Berry Anthocyanins (Blackcurrant, Blueberry, Cherry, Cranberry, Grape)
As shown in Table 2
, berries contain a range of different flavonoid subclasses, but they are typically richest in anthocyanins. Initial berry studies predominantly investigated the cognitive effects of whole fruit. For example, Dodd [25
] demonstrated improved accuracy on a letter memory task (measuring working memory) following freeze dried whole blueberries (200 g fresh equivalent, 631 mg anthocyanidins), in 19 young adults at a postprandial time point of 5 h (d
= 0.57). The study employed a double blind, crossover design with an energy matched control condition. No effects were observed at an earlier time of 2 h or for other measures of executive function, memory, or mood. For a subset of participants, blood samples taken 1 h postprandially revealed a trend towards increased plasma levels of brain-derived neurotrophic factor (BDNF) in the blueberry condition. Unfortunately, cognition was not measured at this time point so it is impossible to say whether the neurochemical changes are related to the cognitive outcome. In the same study, older adults’ BDNF values decreased from baseline for both blueberry and placebo conditions, but the decrease at the 1 h time point was attenuated in the blueberry condition. These older adults (n
= 18) showed improved performance on an immediate word recognition task at both 2 h (d
= 0.44) and 5 h (d
= 0.69) postprandially, but no improvements in executive function or mood were observed.
A later study by Whyte and Williams [26
], using fresh whole blueberries (200 g, 143 mg anthocyanins), investigated cognitive effects in children. They found no effects at 2 h for a range of executive function tasks, but did observe a significant improvement in delayed word recall using the Rey auditory verbal learning task (RAVLT) (d
= 0.74). This was a small, crossover study with only 14 participants. As no baseline measures were taken, variations in performance across test days may have reduced the statistical power. Nevertheless, the medium effect size for the RAVLT provides good evidence for positive effects of blueberry flavonoids in children. Whyte, Schafer and Williams [27
] conducted a larger (n
= 21) double-blind, placebo-controlled, crossover study investigating the cognitive effects of 2 separate blueberry doses (127 mg and 253 mg anthocyanins), again in children. The highest dose resulted in significant improvements in immediate word recall after 1.25 h (d
= 0.80), and in delayed word recognition after 6 h (d
= 0.78). Improved accuracy was observed during a flanker interference task after 3 h, although only for cognitively demanding incongruent trials (d
= 0.78). However, reaction times for a Go-NoGo measure of inhibition revealed significantly faster performance following the placebo compared with the blueberry interventions.
The positive blueberry effects in older adults and children appear to be focussed on episodic memory, whereas improvements in executive function are more consistent in young adults. The differences in cognitive domains may be an artefact of the small sample sizes, but could also be indicative of age differences in the capacity for improvement in underlying neuronal structures. For example, hippocampal function may be more receptive during development in childhood and decline in old age, whilst frontal regions associated with executive function may be more sensitive in young adulthood. It is noteworthy that neurochemical changes in BDNF were apparent after 1 h, yet distinct time points for memory effects emerged at 1.25–2 h and 5 h, but not at an intervening 3 h time point. At this stage only BDNF trends have been observed, and not directly in association with cognitive changes. Although it is perhaps premature to comment on the relationship between acute changes in cognition and BDNF, it has nevertheless been posited that flavonoid induced increases in BDNF may facilitate stronger memory encoding [27
]. Possible mechanisms of action are discussed below. Overall, the timings of cognitive effects are likely to be related to the digestion, absorption and metabolism of flavonoids, but further mapping of cognitive and physiological observations is required in order to resolve inconsistencies within the current observations.
The flavonoid content of blueberries is known to vary widely depending on growing, processing and storage conditions [28
]. The same 200 g quantity of whole blueberries used in the first two studies [25
] described above showed extreme differences in flavonoid content (631 mg anthocyanidins and 143 mg anthocyanins respectively). This highlights the importance of analysing fresh fruits for their flavonoid content when conducting an intervention. It is also important to note that compositional analysis of anthocyanins typically (but not exclusively) involves the removal of saccharide conjugates prior to quantification; therefore anthocyanin content is often reported as anthocyanidin equivalent. This difference is critical when comparing doses between studies. For example, a berry intervention reported to contain 100 mg cyanidin may actually contain 156 mg chrysanthemin (a saccharide of cyanidin). Some studies reviewed here appear to use the terms anthocyanins and anthocyanidins interchangeably without acknowledging this distinction, making it unclear whether a reported anthocyanin dose is actually referring to an equivalent anthocyanidin dose.
Similarly to blueberries, blackcurrants are a rich source of anthocyanins. Watson et al
] conducted a double-blind, controlled crossover trial of two blackcurrant extracts (cold-pressed juice or freeze-dried powder). Improved attention compared with an energy-matched control was observed in 36 young adults during a 70-min-long, cognitively fatiguing battery, beginning 1 h postprandially. Specifically, declining accuracy on a rapid visual information processing (RVIP) task was attenuated after taking the powdered extract (d
= 0.10, d
= 0.47, d
= 0.47, d
= 0.49, d
= 0.49, d
= 0.59, d
= 0.56, measured for 7 task repetitions; once every 10 min). Similarly, a slowing of reaction time on a digit vigilance task following both blackcurrant and placebo interventions was attenuated after taking the juiced extract (d
= 0.60, d
= 0.73 and d
= 0.60 for the 1st, 4th and 7th repetitions of the task, respectively). No effects were observed for the Stroop test (a measure of inhibition and attention), or for subjective measures of mood and mental fatigue. The total polyphenol content of the two extracts were matched at 525 mg/60 kg bodyweight. However, the anthocyanin content differed slightly; 483 mg/60 kg bodyweight for the powder and 467 mg/60 kg for the juice. This difference was also reflected in the analysis of plasma anthocyanin levels, which were observed to be higher following consumption of the powdered extract compared with the juice. Interestingly, the juice but not the powder was observed to inhibit platelet monoamine oxidase (MAO) and to attenuate blood glucose decline over the duration of the 70-min task battery. This study suggests that the way an extract is prepared may influence cognitive and physiological outcomes, however as different blackcurrant cultivars were used for each extract, the contrasting observations may simply represent compositional differences such as the ratio of flavonoid subclasses present. For the juiced blackcurrant extract, MAO inhibition and blood glucose regulation emerge as possible mechanisms of action further to the neurochemical changes observed for blueberries. The significant cognitive effects were observed for tests of executive function (RVIP and to some extent vigilance) which is consistent with the executive function benefits reported in healthy young adults following blueberry anthocyanins.
In a double-blind crossover intervention study, Hendrickson and Mattes [31
] investigated whether an acute dose of grape juice would mitigate deficits in mood and cognition that commonly occur following a large meal. Approximately 600 mL (10 mL/kg), containing around 580 mg anthocyanins, was administered to young adult smokers along with a standardised lunch. Smokers were selected on the rationale that this population have an increased propensity to oxidative stress, and because smoking abstinence can exaggerate the post-meal dip in cognitive or affective state, thus this population may be more sensitive to the effects of flavonoids than healthy non-smokers. This was a large study (n
= 35) with considerable statistical power, yet no significant effects of grape juice were observed 1 h postprandially when compared to an energy matched placebo condition. Mood ratings for positive mood states (pleasure, arousal and vigor) were observed to decline under both grape and placebo conditions, similarly ratings of negative mood states (confusion and fatigue) increased under both conditions. Although mood generally declined, word fragment completion task performance did not significantly change over time in either condition. It is unfortunate that performance on only one cognitive domain was examined (implicit memory), which is an area that has not previously been considered with respect to flavonoid intervention. Studies may be more likely to observe effects on traditional measures of explicit memory and executive function.
Recently, Caldwell et al
] published their investigations into the effects of cherry flavonoids. Following administration of 300 mL cherry juice (55 mg anthocyanins) to younger adults (n
= 6), older adults (n
= 5) and older adults with mild cognitive impairment (n
= 5), tests of executive function, speed of processing, and verbal learning and memory were performed at baseline and 6 h postprandially. At 6 h, the older adults displayed improved task switching performance compared to baseline (db
= 0.75). No other cognitive effects were observed. The authors attribute this single effect to type 1 error, citing attrition of participants in that group as a likely cause. However the small sample size in all groups suggests that the whole study is likely to be severely underpowered. The lack of an energy matched, low flavonoid control condition is also cause for concern; a second crossover condition only administered the same juice in three separate 100 mL aliquots each consumed 1 h apart. No cognitive effects were observed relative to baseline following consumption of the juice in these consecutive smaller doses. A further problem may be the intervention itself; the anthocyanin content of the cherry juice appears very low compared to some of the above studies. A considerably larger, controlled study is therefore needed to determine if cherry anthocyanins elicit acute cognitive effects similar to those of other anthocyanin-rich fruits.
3.1.2. Citrus Hesperidin (Orange)
As with berries, citrus fruits contain several different flavonoids, but they are richest in the flavanone hesperidin. In a crossover intervention study, Lamport et al.
] supplemented 24 young adults with a commercially available flavanone-rich orange juice (70.5 mg total flavonoids; 42.15 mg hesperidin, 17.25 mg naringin, 6.75 mg narirutin) or an energy matched control. An extensive battery of tasks including measures of vision, episodic memory, processing speed, working memory, and other executive functions were performed 2 h postprandially. Significant improvements when compared to both control and baseline were observed only for the digit symbol substitution task (DSST), a measure of psychomotor processing speed (d
= 0.30). An additional group of participants underwent fMRI assessment using arterial spin labelling (ASL). Increased cerebral blood flow (CBF) was observed at 2 h but not 5 h postprandially. However, as cognition was not measured directly in conjunction with CBF (and not at all at 5 h) it remains unclear whether the observed improvements in processing speed are causally related to the CBF changes. The flavonoid dose used in this study was low, particularly when compared with the typical doses of other flavonoid subclasses reviewed here. If flavonoid effects are dose dependent then a higher dose may have elicited increased cognitive benefit. Indeed, a subsequent study by Alharbi et al.
] used orange juice fortified with additional flavanone rich orange pulp, to achieve a greater total flavonoid content (272 mg; 220.46 mg hesperidin, 34.54 mg narirutin). Middle aged adults (n
= 24, 30–65 years) showed improved psychomotor performance on a finger tapping task at both 2 h (d
= 0.87) and 6 h (d
= 0.62). Improvements in attention and general executive function, as measured by a continuous performance task (CPT), were observed at 6 h (d
= 0.58). The orange juice was observed to attenuate a decline in subjective alertness throughout the testing period compared to a sugar-matched control.
From this limited research, orange juice appears to benefit psychomotor performance across age groups at a time point also associated with increased CBF. Higher doses appear to elicit greater effect sizes for psychomotor performance and provide additional benefits in executive function, suggesting that flavanone effects are dose-dependent.