Omega-3 Polyunsaturated Fatty Acids and Behavior

Humans typically consume about 20 different types of fatty acids in the diet, which can be grouped as either saturated or unsaturated fatty acids. Saturated fatty acids have single bonds between the carbon atoms and are rigid in nature. Unsaturated fatty acids may have one (monounsaturated) or more (polyunsaturated) double bonds and the position of the first double bond in relation to the omega end determines whether a polyunsaturated fatty acid is termed an omega-3 (n-3) or an omega-6 (n-6) fatty acid. Mammals are capable of manufacturing every fatty acid required for biological processes except for two; namely linoleic acid (LA, n-6) and a-linolenic acid (ALA, n-3). These are termed the “essential” fatty acids and must be acquired via the diet (Simopoulos 2000). LA and ALA are sometimes referred to as “parent” fatty acids as it is from these that their respective long-chain biologically active metabolites are derived. Arachidonic acid (AA, n-6) is the major metabolite of LA, whereas eicosapentaenoic acid (EPA, n-3) and docosahexaenoic acid (DHA, n-3) are the major metabolites of ALA. AA, EPA, and DHA are synthesized from their respective precursor parent fatty acids by a series of elongations and desaturations that, despite the fact that the conversion pathways for n-6 and n-3 fatty acids are entirely independent, require the same enzymes at each step. There is also some evidence to suggest that DHA can be “retro-converted” into EPA, although rates of only 20% have been observed (Gronn et al. 1991). The metabolism of LA and ALA is predominantly carried out in the endo-plasmic reticulum of the liver, in certain structures in the central nervous system such as glial cells (Moore 2001) and the choroid plexus vasculature (Bourre et al. 1997), and has also been observed at low rates in the placenta (Haggarty 2004).

The consumption of n-3 PUFAs has been falling gradually over the past 100–150 years; the typical “Western” diet of today is characterized by a marked decrease in overall fish consumption and increased intake of n-6 PUFAs that are abundant in cooking oils and processed foods (Simopoulos 2008). There is evidence to suggest that humans evolved on a diet where n-6 and n-3 PUFAs were consumed in approximately equal amounts (1–4:1) (Simopoulos 1991), whereas the consumption ratio of n-6 and n-3 PUFAs in the current Western diet is estimated anywhere between 10: and as much as 25:1 (Simopoulos 2000). There is also mounting evidence to suggest that decreased dietary intake of n-3 PUFAs, DHA and EPA in particular, is a risk factor for a plethora of different diseases including cardiovascular disease (Mori and Woodman 2006), inflammatory disease (De Caterina and Basta 2001), and many neurodevelopmental and psychiatric conditions such as attention-deficit hyperactivity disorder (ADHD), dyslexia, depression, schizophrenia, and dementia (Bourre 2005). It follows that for these two n-3 PUFAs to be implicated in such a range of seemingly unrelated conditions, they are likely to influence fundamental processes common to most cells.

Indeed, once consumed (or metabolized) DHA and EPA are incorporated at the sn-2 position of cellular membrane phospholipids in every type of tissue, where they compete for incorporation at the same position with AA (Calder 2006a). Under certain conditions, DHA and EPA (and AA) are released from the cell membrane by the action of several phospholipases (Farooqui et al. 1997), where they are metabolized further to form potent secondary signaling molecules classed as either eico-sanoids (from EPA) or docosanoids (from DHA) (Tassoni et al. 2008). The dietary intake of n-3 PUFAs is, therefore, reflected in the composition of all cell membranes, which can impact a number of varied cellular processes, described in the following.

Communication between neurons relies on the exchange of ions across the cellular membrane, with maximum efficiency occurring at an “optimal” value where the physical state of the membrane is neither too rigid nor too fluid (Yehuda et al. 1999). The structure of the cell membrane varies greatly, depending on the fatty acids that make up the hydrophobic “tail” of the phospholipids. For example, rigid saturated fatty acids allow phospholipids to pack tightly together, whereas the insertion of double bonds along the hydrocarbon chain alters the properties of the fatty acid. Therefore, as the degree of unsaturation increases, the chain becomes more flexible and starts to “kink.” DHA, which has six double bonds and is preferentially incorporated at the sn-2 position of the phospholipids phosphatidylethanolamine and phosphatidylserine, in particular, can adopt countless looped and helical conformations and, thus, tight pack-ing of these DHA-rich phospholipids is prevented, consequently increasing the fluidity

of the membrane (Feller et al. 2002). EPA, possessing five double bonds can also adopt multiple conformations, but the extra double bond present in DHA renders this fatty acid unique and highly specialized, as evidenced by its high density in selected tissues (Stillwell and Wassall 2003). More specifically, DHA is heavily concentrated in the cerebral frontal cortex of mammals and comprises anywhere between 10% and 20% of total fatty acids of the brain (McNamara and Carlson 2006) and represents around 30%–40% of the PUFAs found in the retinal rod outer segment (Makrides et al. 1994). Modulation of membrane fluidity in these tissues occurs with dietary manipulation of n-3 PUFAs (Connor et al. 1990; Anderson et al. 2005), and variations in concentrations of n-3 PUFAS in the cell membrane have been shown to impact a number of different cellular processes, all of which have the potential to impact upon brain function and hence behavior. For example, both DHA and EPA have been shown to affect the activities of membrane bound enzymes (e.g., Slater et al. 1995; Turner et al. 2003), ion channels (e.g., Kang and Leaf 1996; Xiao et al. 1997; Seebungkert and Lynch 2002), and gene expression (e.g., Kitajka et al. 2002; Barcelo-Coblijn et al. 2003), which can in turn influence signal transduction and neuronal transmission. In addition, levels of dopamine (Zimmer et al. 2000a), serotonin (de la Presa Owens and Innis 1999), and acetylcholine (Aid et al. 2003) have been observed to either increase or decrease fol-lowing either an n-3-enriched or n-3-deficient diet. Further to this, DHA in particular has been shown to have a number of neuroprotective properties. These include pre-venting apoptosis when DHA is metabolized into phosphatidylserine (Kim et al. 2000) and reducing oxidative stress (Mori et al. 2000). In addition, the docosanoid deriva-tives of DHA, described later, have also been shown to be neuroprotective.

Cell membrane incorporation of DHA and EPA also has an effect on the production of two classes of secondary signaling molecules, namely eicosanoids or docosanoids. These molecules are powerful biological compounds responsible for mediating many aspects of the inflammatory response (Calder 2006a). Eicosanoids—catego-rized further as either leukotrienes, thromboxanes, or prostaglandins—can also be derived from AA upon its release from the cell membrane, and tend to be more potent and pro-inflammatory than those originating from EPA (Schmitz and Ecker 2008). However, higher intake of dietary EPA leads to increased incorporation of these molecules into membrane phospholipids in a dose response manner and at the expense of membrane incorporation of AA (Calder 2007). Consequently, there is a shift away from production of pro-inflammatory, vaso-constricting, and platelet-aggregating AA-derived eicosanoids, and an increase in the production of anti-inflammatory EPA-derived ones (Gibney and Hunter 1993).

Like eicosanoids, docosanoids are chemical signaling molecules, produced via con-trolled oxidative degeneration of DHA within or adjacent to the cell membrane (Kidd 2007). Three classes of docosanoids have been identified—docosatrienes, resolvins, and protectins—and have been shown to have neuroprotective qualities. The novel neuroprotectin D1 (NPD1) has been shown to attenuate apoptosis in the presence of oxidative stress and provides protection to neuronal cells in animal models of brain ischemia and neurodegeneration (reviewed in Bazan 2006). More specifically, in Alzheimer’s disease (AD) rat models NPD1 repressed the expression of pro-inflammatory 0-amyloid-activated genes. Moreover, the recently discovered E-series and D-series resolvins, derived from EPA and DHA, respectively, have also been identified as having anti-inflammatory properties that are not related to altering lipid mediator profiles (i.e., inhibited production of AA-derived eicosanoids), but by inhibiting the expression of pro-inflammatory cytokine genes such as nuclear factor ic B and/or per-oxisome proliferator–activated receptor (Calder 2006b). Taken together, the modulation of eicosanoid and docosanoid production is one potential mechanism by which dietary DHA and EPA could prevent the occurrence or ameliorate the symptoms of inflammatory diseases linked to n-3 PUFA intake, including depression (Das 2007), ADHD (Richardson 2006), schizophrenia (Yao and van Kammen 2004), AD (Pratico and Trojanowski 2000), atherosclerosis (von Schacky 2000), rheumatoid arthritis (Kremer 2000), inflammatory bowel disease (De Caterina et al. 2000), and possibly some bronchial diseases such as asthma (Belluzzi et al. 2000).

A final function of n-3 PUFAs relates to their effects on various aspects of cardiovascular function. Given that cerebrovascular events are a risk factor for neurodegenerative, along with the fact that the cardiovascular system is responsible for the delivery of nutrients to the brain, it follows that any compound that modulates cardiovascular parameters could exert a secondary effect on brain function and behavior. Indeed, a number of different cardiovascular parameters have been shown to be modified by dietary n-3 PUFAs including increased arrhythmic threshold via modulation of sodium and calcium ion channels (Kang and Leaf 1996), decreased platelet aggregation (Mori et al. 1997), lowered triglycerides (Nestel 2000), lowered blood pres-sure (Morris et al. 1993; Geleijnse et al. 2002), and improved arterial and endothelial function via increased nitric oxide synthesis (Harris et al. 1997; Armah et al. 2008).

In summary, DHA and EPA are involved in a number of varied fundamental functions at the cellular level. In the brain, DHA is heavily enriched in the cerebral cortex where its incorporation into the phospholipid bilayer of neural cell membranes confers optimal membrane fluidity, resulting in improved membrane function as regards signal transduction and neurotransmission. Furthermore, there is evidence to suggest that the expression of a number of genes and the production of various neurotransmitters is sensitive to dietary intake of n-3 PUFAs, suggesting a role for n-3 PUFAs in these processes. In addition, the DHA and EPA incorporated into cell membranes throughout the body can be subsequently released and metabolized further to produce potent secondary signaling molecules that are essential in the resolution of the immune response and may also be neuroprotective. Finally, dietary n-3 PUFAs modulate a number of cardiovascular parameters, which may contribute to reduced risk of cardiovascular events. Given the fundamental nature of n-3 PUFAs and DHA and EPA in particular, it is plausible that alterations in dietary intake could potentially impact upon brain function and behavior. The following section reviews the current literature on the behavioral effects of n-3 PUFAs in animals and humans.

Our knowledge of the impact dietary n-3 PUFAs have upon cognitive function has been greatly extended by the investigation of their effects in animals, the majority of which have been conducted using rodents. Overall, the evidence from these studies indicates that carefully controlled n-3-deficient diets lead to a decrease in levels of brain DHA, which is associated with poorer performance on a selection of learning and memory tasks such as Morris Water Maze (Moriguchi et al. 2000; Fedorova and Salem 2006), avoidance learning (Garcia-Calatayud et al. 2005), and olfactory discrimination tasks (Greiner et al. 2001). In addition, third-generation rats (87% reduction in brain DHA) have been found to perform worse than second-generation rats (83% reduction in brain DHA) (Moriguchi et al. 2000). Interestingly, in both sets of animals, performance was inversely related to levels of docosapentaenoic acid (DPA, n-6) in the frontal cortex, suggesting that the reciprocal replacement of DHA with DPA has significant consequences.

In older rats, impairments in tasks that involve complex motor skills and spatial memory decline throughout the lifespan (Shukitt-Hale et al. 1998), which may be attributable to the observed reductions in brain lipids, have been consistently observed in aged animals (e.g., Ulmann et al. 2001). Long-term potentiation (LTP), commonly thought to be the biological process underlying learning and memory, is reduced in aged rats (Landfield et al. 1978). In addition, both AA and DHA are significantly decreased in these animals (McGahon et al. 1999). Interestingly, the ability of rat hippocampal dentate gyrus cells to sustain LTP is negatively correlated with the concentration of both AA and DHA in these cells, suggesting a link between the prevalence of long-chain PUFAs and learning and memory (McGahon et al. 1999). Eight weeks of n-3 PUFA supplementation (10 mg/day DHA) is sufficient to restore membrane DHA, which is accompanied by a reversal of the deficits in the ability to sustain LTP (McGahon et al. 1999). Other studies have shown that DHA supplementation can restore radial arm maze task performance in both n-3-deficient (Gamoh et al. 2001) and n-3-adequate (Carrie et al. 2000) aged rats. Together these investigations in aged animals suggest a theoretical basis for and observable benefit of n-3 PUFA supplementation in reducing or reversing age-related impairments.

In humans, n-3 PUFA deficiency to the extent that is observed in animals is extremely rare and only a handful of cases have ever been reported, most commonly as the result of administration of total parenteral nutrition (feeding exclusively via intra-venous drip) containing very little or no ALA. Rough, dry skin and hair, excessive thirst and abnormal vision are common features of this type of deficiency; symptoms can be reversed once ALA is reintroduced to the diet (Holman et al. 1982). n-3 PUFA status can be determined in humans by measuring the concentrations of ALA, DHA, and EPA in peripheral tissues such as serum/plasma or erythrocytes. By comparing the n-3 status of healthy normal volunteers to those of various patient groups, it has been revealed that individuals diagnosed with several neurodevelopmental disorders such as ADHD and autism (Bell et al. 2000; Burgess et al. 2000; Schuchardt et al. 2009), along with a number of psychiatric conditions including depression (Edwards et al. 1998), schizophrenia (Assies et al. 2001), and AD and dementia (Conquer et al. 2000), have significantly lower levels of n-3 PUFAs. Collectively, these findings again suggest that adequate intake and incorporation of n-3 PUFAs is a requirement for normal functioning. The results from studies that have used n-3 supplementation as treatment for symptoms of these conditions have been mixed, however, and further investigation is required. In the next section the role of n-3 PUFAs in a number of neuropsychiatric and developmental conditions is outlined, along with an evaluation of the current evidence of their use in the treatment of these conditions. The section will end with a review of the current knowledge of the effects of n-3 PUFA supplementation on behavioral outcomes in healthy individuals.

About 10%–20% of postpartum women are diagnosed with postpartum depression (PPD). As maternal stores of fatty acids are depleted during pregnancy to ensure an adequate supply for central nervous system development of the growing neonate, some researchers have explored the hypothesis that without sufficient dietary intake of fatty acids, mothers may increase their risk of suffering from PPD (Holman et al. 1991). In rats, it has been observed that an inadequate supply of dietary DHA is enough to result in a 21% decrease in brain DHA in just one reproductive cycle (Levant et al. 2006), but the extent and possible consequences of depletion in humans has yet to be established. In a cross-national study, Hibbeln (2002) discovered that seafood intake and levels of DHA in breast milk were inversely associated with depressive symptoms as measured by the Edinburgh Postnatal Depression Scale (EPDS) in 22 countries world-wide, but another study of 80 new mothers found no relationship between postnatal n-3 fatty acid status and postnatal depression (Browne et al. 2006). In addition, the results from the few intervention trials that have been conducted in this population generally do not support n-3 PUFAs as a treatment of PPD, although large RCTs are still required. In a small open-label trial, supplementation of 2.96 g/day DHA and EPA starting at between 34 and 36 weeks’ gestation did not prevent PPD in four out of seven participants (Marangell et al. 2004). Freeman and colleagues have conducted two intervention trials in women who have been diagnosed with depression following birth. The first of these studies was an open-label pilot trial where participants (N = 15) received approximately 1.9 g/day EPA + DHA for 8 weeks (Freeman et al. 2006a). Authors reported a 40.9% decrease in depressive symptoms on the EPDS but in a second randomized dose-ranging study where treatments ranged from 0.5 to 2.8 g/day as adjunctive treat-ment to supportive psychotherapy, the authors found no difference between groups, with all groups reporting reduced scores on the EPDS and Hamilton Depression Rating Scale (Freeman et al. 2008). It is possible that the association between maternal intake of n-3 PUFAs and PPD has been overestimated; results from the Danish National Birth Cohort, a large prospective study, reveal little evidence to support a link between maternal fish and n-3 PUFA intake and rates of PPD (Strom et al. 2009). A review of the extant evidence in this area concluded that the results are not conclusive overall, but do warrant further investigation (Borja-Hart and Marino 2010).

Dietary n-3 PUFAs have also been implicated in other neuropsychiatric conditions such as bipolar disorder (BD) and schizophrenia. The similarities between the effects of mood stabilizers such as lithium and valproate—commonly used in the

treatment of BD—and DHA and EPA, on the enzyme protein kinase C (PKC) have led researchers to consider n-3 PUFAs as an alternative to standard pharmacological treatment for BD. Further, epidemiological studies have revealed an inverse relation-ship between seafood consumption and lifetime prevalence rates of BD (Noaghiul and Hibbeln 2003). However, evidence from intervention trials is inconclusive, with some published trials reporting a benefit of n-3 PUFAs (Stoll et al. 1999; Osher et al. 2005; Sagduyu et al. 2005; Frangou et al. 2006), while others do not (Marangell et al. 2003; Keck et al. 2006). A systematic review of the extant literature in this area concluded that although n-3 PUFAs are well tolerated by patients with BD and the evidence seems to show an association between n-3 use and symptom reduction, further studies are required in order to confirm their efficacy in the treatment of BD (Turnbull et al. 2008).

A similar pattern of findings is observed in schizophrenia. The popular “dopamine hypothesis” of schizophrenia proposes that negative symptoms (flat affect) result from reduced activity of the dopamine systems in the prefrontal area, and positive symptoms (delusions and thought disorder) from increased activity of the dopamine systems in the limbic system (Davis et al. 1991). This theory can explain the relationship between dopamine kinetics and the psychiatric symptoms of schizophrenia, but fails to address the cause of the abnormal activities of dopaminergic neurons (Ohara 2007). Zimmer and colleagues discovered that rats who had been fed an n-3-deficient diet suffered a reduction in the number of presynaptic dopamine vesicles and also that basal dopamine metabolism was increased (Zimmer et al. 2000a,b). Dietary n-3 deficiency has also been shown to reduce the number of D2-receptors in the frontal lobe in both rats (Delion et al. 1994) and piglets (de la Presa Owens and Innis 1999). It has also been observed that compared to controls, schizophrenia patients have lower levels of plasma n-3 PUFAs (Assies et al. 2001). Therefore, in an attempt to integrate all of the evidence, Ohara (2007) proposed that the n-3 PUFA abnormalities found in schizophrenia stem from the dysfunction of the enzyme phospholipase A2 (PLA2). It follows that increased activation of PLA2 observed in patients suffering from schizophrenia may cause the excessive depletion of PUFA from the sn-2 position of cell membrane phospholipids in the body and brain. Dopamine concentration, the number of dopamine vesicles, and the number of D2 receptors are decreased in the prefrontal presynaptic terminals (resulting in the negative symptoms) and these decreases have a knock-on effect for the limbic dopamine system (resulting in the positive symptoms) (Ohara 2007).

Despite the apparent plausibility of this integrated theory, a Cochrane review of PUFA supplementation in schizophrenia concluded that data from the six trials that met the inclusion criteria were inconclusive, and the value of treating schizophrenia with PUFA remains unfounded (Joy et al. 2006). This conclusion was formed largely on the basis that of the six trials, only one enrolled more than 100 participants (Peet and Horrobin 2002) and in only one study did the intervention period exceed 3 months (Fenton et al. 2001). Neither of these studies produced compelling evidence to support the use of n-3 in the treatment of schizophrenia. Only large, longitudinal RCTs will be able to provide sufficient evidence as to whether n-3 PUFAs have a clinically significant and positive impact in the treatment of this illness.

AGE-RELATED COGNITIVE DECLINE AND DEMENTIA

Cognitive function naturally declines with age and has been attributed to a num-ber of factors including reduced synaptic plasticity, decreased membrane fluidity, and increased oxidative damage (Willis et al. 2008). There is growing evidence, however, that various lifestyle factors can either promote or attenuate cognitive aging. These include smoking (Swan and Lessov-Schlaggar 2007), alcohol consumption (Peters et al. 2008), exercise (Colcombe et al. 2003), and diet (Del Parigi et al. 2006; Barberger-Gateau et al. 2007). In particular, one of the dietary factors that have been explored in detail is intake of fatty acids. For example, the Dutch prospective population-based Zutphen Elderly Study identified that LA was positively associated with cognitive decline over a 3 year period (defined as a >2 point drop in Mini Mental State Examination) in 476 men aged 69–89 years (Kalmijn et al. 1997). A recent reanalysis of the same data was able to identify that in this sample of elderly men, those who did not eat fish observed a 1.2 point decline in MMSE score at the 5 years follow-up, as opposed to only a 0.3 point decline in men who reported eating fish (van Gelder et al. 2007). Additionally, a cross-sectional study by the same group identified that oily fish consumption (measured using a FFQ) was significantly associated with a reduced risk of global cognitive function impairment and psychomotor speed in participants of 45–70 years, independent of other confounding factors (e.g., age, sex, education, smoking, alcohol consumption, energy intake) (Kalmijn et al. 2004).

Findings from the Chicago Health and Aging Project (CHAP), conducted in 2560 participants aged 65 years and older over a period of 6 years, also discovered that fish intake was associated with a slower rate of cognitive decline at the 6 years follow-up. More specifically, among those who consumed one fish meal per week, decline was 10% slower than those who consumed fish less than weekly and 13% slower for those who consumed two or more fish meals per week, adjusted for age, sex, race, education, cognitive activity, physical activity, alcohol consumption, and total energy intake. What the authors could not conclude is whether it was n-3 PUFAs that were the relevant dietary constituent in fish accountable for this finding (Morris et al. 2005). The prospective population-based Etude du Vieillissement Ateriel (EVA) study evaluated fatty acids in erythrocyte membranes and performance on the MMSE in a sample of 246 63–74 year olds (Heude et al. 2003). These authors found that higher proportions of stearic acid (a saturated fatty acid) and total n-6 PUFAs (LA, AA, y-linolenic acid (GLA), DPAn-6) were associated with greater risk of cognitive decline and that a higher proportion of total n-3 PUFAs (ALA, DHA, EPA, DPAn-3) was associated with a lower risk of cognitive decline over a 4 year period. Similarly, intake of EPA and DHA (estimated via a food frequency questionnaire) was inversely associated with cognitive impairment (MMSE). Finally, higher plasma n-3 PUFA proportions in a sample of 807 healthy participants aged 50–70 years predicted less decline in sensorimotor speed and complex speed over a 3 year period, although there were no associations between n-3 PUFA proportions and memory, information processing speed or word fluency, and no significant associations were detected at baseline between n-3 status and performance in any of the five assessed cognitive domains (Dullemeijer et al. 2007).

It is only recently that data from large-scale prospective randomized intervention trials evaluating the effects of n-3 PUFAs on cognitive function in older adults have been available; however, results from these trials have been conflicting. The OPAL (Older People And n-3 Long-chain polyunsaturated fatty acids) study assessed the effects of a daily fish oil supplement containing 0.5 g DHA and 0.2 g EPA on cognitive performance on the California Verbal Learning test and other measures of memory and attention in 867 men and women aged 70–79 years (at baseline), but did not find any significant effects of the treatment. Similarly, the 26 weeks intervention trial in 302 healthy older adults reported by van de Rest et al. (2008) also did not find any effects of either a high (1.8 g EPA + DHA) or lower dose (0.4 g EPA + DHA) com-pared to placebo on a range of cognitive assessments. On the other hand, the memory improvement with docosahexaenoic acid study (MIDAS) intervention trial found a significant effect of 24 weeks supplementation with 0.9 g DHA on 485 healthy adults (=55 years) who were classified as having age-related cognitive decline (ARCD) on learning and episodic memory tasks, but not working memory or executive function tasks (Yurko-Mauro et al. 2010). This latter study may have potentially highlighted a subgroup of healthy older adults in which administration of n-3 PUFAs has benefi-cial effect. Further research would need to confirm this hypothesis.

The progression of ARCD to cognitive impairment is rising dramatically the world over and currently around 24.2 million people are affected by dementia, with 4.6 million new cases reported each year; AD accounts for about 60% of cases (Ferri et al. 2005). A number of observational studies in humans have examined the rela-tionship between intakes of n-3 PUFAs, as measured by various food frequency questionnaires (FFQ), and diagnosis of dementia or AD, but overall the results are conflicting. Barberger-Gateau et al. (2002) found in their analysis of the PAQUID epidemiological study (N = 1674 aged 68 years or more) that those participants who consumed fish or seafood at least once a week were at a lower risk of developing dementia, including AD at the 7 year follow-up. However, after adjusting for education level, which was positively correlated with fish intake, the strength of the association diminished somewhat. A publication from the CHAP cohort demonstrated, after a mean follow-up of 3.9 years, that a higher intake of DHA and weekly fish consumption reduced the risk of AD, although EPA was not associated with a reduced risk (Morris et al. 2003). Conversely, results from the prospective population-based Rotterdam study (N = 5395) found no association between n-3 intake and risk for any type of dementia (Engelhart et al. 2002). Similarly, results from the Canadian Study of Health and Aging also do not suggest that an association between total n-3 PUFAs, DHA, or EPA and incidence of dementia or AD (Kroger et al. 2009). In addition, the results from two other large-scale studies that initially indicated an inverse association between n-3 PUFAs and incidence of AD and dementia were attenuated once sex, age, and education were adjusted for (Huang et al. 2005; Schaefer et al. 2006).

Despite these mixed reports, the biological basis for pursuing research in this area is compelling; n-3 PUFAs possess three properties by which they may protect against the development of dementia, which include increasing cerebral blood flow, attenuating inflammation, and reducing amyloid production (reviewed in Fotuhi et al. 2009). Results from animal studies are indeed encouraging; in their review of the protective effects of n-3 PUFAs in AD, Boudrault et al. (2009) conclude that treatment with

DHA in rodent models of AD consistently protects against the development of AD, with a number of observable effects in the brains of animals fed DHA compared to controls including decreased pro-apoptotic proteins and secretion of amyloid beta (AP) and increased activity in the PI-3 kinase cascade, a neuroprotective pathway shown to be reduced in AD. Coupled with these physiological changes are studies showing improvements in cognitive function. One group from Japan have focused particularly on this issue, and have consistently shown protective effects of n-3 PUFA administration on spatial learning ability in Ap-infused rats (Hashimoto et al. 2002; Hashimoto et al. 2005a,b, 2008). However, it is worth noting that the quantity of n-3 PUFAs given to these animals is two to four times greater than the current intake in humans (Boudrault et al. 2009). Interestingly, in humans, levels of DHA in the brains of AD patients do not significantly differ from those that are normal, although levels of stearic acid (frontal and temporal cortex) and AA (temporal cortex) are reduced, and oleic acid is increased (frontal and temporal cortex), indicating some differences in brain fatty acid composition (Fraser et al. 2009). Compared to animal studies, intervention trials in humans, however, have not been met with the same success. A dose-ranging intervention in 302 participants aged 65 years or older with an MMSE score of >21 found no effect of either dose of fish oil containing either 400 or 1800 mg DHA + EPA on cognitive function (memory, sensorimotor speed, attention, executive function) compared with placebo following 26 weeks of dietary supplementation (van de Rest et al. 2008). Similarly, the OmegaAD clinical trial examined the effects of n-3 PUFA supplementation in 174 patients with mild to moderate AD. In this one-way crossover trial, the active treatment consisted of daily dietary supplementation with 1.6 g of DHA and 0.6 g EPA. At 6 months there was no difference between groups on either the MMSE or the AD Assessment Scale. However, in a subgroup of participants with very mild cognitive dysfunction there was a significant reduction in MMSE decline rate, and this was replicated in the crossover group at 12 months (Freund-Levi et al. 2006). These authors also suggest that in terms of the neuropsychiatric symptoms of AD, carriers of the APOE4 gene might be more susceptible to the effects of treatment with n-3 PUFAs, although this is an avenue of investigation that needs to be pursued further (Freund-Levi et al. 2007). Lim et al. (2006) conclude in their Cochrane review that there is a growing body of evidence from biological, observational, and epidemiological studies suggesting a protective effect of n-3 PUFAs against dementia. The level of this effect remains unclear, however, and, to date, dietary recommendations in relation to fish and n-3 PUFA consumption and risk of dementia cannot be made. It is hoped that the results of the DHA in Slowing the Progression of AD study, a prospective 18 months intervention trial in 400 participants aged 50 or older with mild to moderate cognitive impairment, could be used to inform the efficacy of n-3 PUFA in the prevention of dementia (Quinn 2007).

Richardson and Ross (2000) were among the first researchers to link neurodevel-opmental disorders such as ADHD, dyslexia, developmental coordination dis-order (DCD), and autism with n-3 PUFA deficiency. These authors noted clinical ommonalities between these conditions such as the preponderance of males that were affected, apparent links between allergies and other immune system disor-ders such as proneness to infections and atopic conditions, abnormalities of mood, arousal and sleep, as well as cognitive impairments in attention and working mem-ory, which suggest disruptions of visual or auditory processing (Richardson 2006). It had also been observed some 25 years previously that individuals with these conditions also shared physical characteristics seen in animals specifically bred on n-3-de-ficient diets such as excessive thirst, frequent urination, rough, dry hair and skin, and follicular keratosis (Colquhoun and Bunday 1981). Indeed, several studies in children with ADHD have demonstrated that these children have lower blood concentrations of PUFAs, namely AA, DHA, and overall concentrations of n-3 PUFAs (Bekaroglu et al. 1996; Stevens et al. 1996; Burgess et al. 2000; Stevens et al. 2003). Given that there is no evidence to suggest that n-3 PUFA intakes are lower in children with ADHD than in healthy children (Ng et al. 2009), the low levels of n-3 PUFAs found in the blood of children with ADHD have been attributed to either inefficient conversion of ALA to EPA and DHA or enhanced metabolism of these fatty acids (Stevens et al. 1995; Burgess et al. 2000). There have been five widely cited intervention trials investigating the effectiveness of n-3 PUFA treatment on symptoms in children with ADHD and related developmental disorders. These studies have varied in design but interestingly the three experiments that report a positive effect of treatment all used a daily treatment regimen lasting 12 weeks or longer and the treatments themselves originated from fish oil, and, therefore, contained both DHA and EPA (Richardson and Puri 2002; Stevens et al. 2003; Richardson and Montgomery 2005). The study by Voigt et al. (2001) found no effect of 345 mg/day DHA for 16 weeks on a wide range of behavioral and computerized measures of ADHD-related symptoms in 54 children diagnosed with ADHD, and Hamazaki and Hirayama (2004) found no effect of treatment on behavioral symptoms of ADHD with a daily fish oil supple-ment for 8 weeks, suggesting the possibility that both the composition of the n-3 PUFA treatment and duration of regimen are key factors in ameliorating symptoms of ADHD and related disorders.

Similarly, a relationship between n-3 fatty acid status and autism has also been demonstrated, although intervention trials showing a pronounced benefit of treat-ment with n-3 PUFAs are lacking. Vancassel et al. (2001) discovered that DHA was decreased by 23% in the plasma phospholipids of autistic children and total fatty acids by 20%. In contrast, a more recent study found that in 16 high-functioning males with autism, DHA and the ratio between total n-3:n-6 PUFAs were increased in plasma phospholipids compared to 22 matched controls, and consequently the authors advised serious caution against treating this condition with n-3 PUFAs (Sliwinski et al. 2006). Despite this, Amminger et al. (2007) published results from a pilot trial wherein they administered seven diagnosed with autistic disorder 7 g/day fish oil for 6 weeks. When compared to matched controls who received a placebo treatment for the same duration, the only significant difference found between groups was on an irritability scale; no differences were found between groups on the social withdrawal, stereotypy, hyperactivity, or inappropriate speech measures. The authors are quick to note the small sample size and the relatively short duration of the trial. No adverse effects on behavior were observed.

n-3 PUFAs have also been linked to dyslexia, and to this end Richardson et al. (2000) examined the associations between the clinical signs of n-3 fatty acid deficiency (excessive thirst, frequent urination, rough, dry hair and skin, etc.) and reading ability, spelling, and auditory working memory in 97 dyslexic children. The authors detected inverse associations between signs of n-3 deficiency and reading and overall ability, and in boys alone, poorer spelling and auditory working memory. This find-ing was reflected in a study of dyslexic adults who filled out two self-report questionnaires; one on signs of fatty acid deficiency and another concerning signs and severity of dyslexia. The authors reported that the signs of fatty acid deficiency were significantly elevated in dyslexic participants and that this reached higher significance in males (Taylor et al. 2000). Cyhlarova et al. (2007) also examined the link between fatty acid status and literacy skills in 32 dyslexic individuals and 20 matched controls. For both groups, better word reading was associated with higher total n-3 concentrations, although it was only in dyslexic participants that a negative correlation was found between reading performance and the ratio of AA:EPA and with total n-6 concentrations, despite there being no significant differences in membrane fatty acid levels between groups, suggesting that, as in ADHD, the ratio of n-6:n-3 PUFAs or a intrinsic disruption in the metabolism of these fatty acids may be a contributing factor in the etiology of these conditions. A collection of preliminary studies reported by Stordy (2000) seems to indicate that impairments of the visual system can be improved with a high-DHA supplement in dyslexic participants, although larger RCTs have yet to be carried out investigating the full extent of the efficacy of n-3 PUFAs in the treatment of dyslexia.

INFANT DEVELOPMENT

The developing fetus requires a supply of both AA and DHA for structural and metabolic functions (Haggarty 2004). The brain and retina require a high concentration of DHA to function optimally and as such, it is thought that the n-3 PUFA com-position of the maternal diet can affect visual and intellectual development (Innis 1991). DHA is deposited in fetal fat stores in the last 10 weeks of pregnancy in the quantity of around 10 g. If the diet is devoid of preformed DHA in the first 2 months of life, then this store is mobilized and would be largely used up, supporting critical developmental processes (Farquharson et al. 1993). While the level of AA in breast milk has been found to remain constant at about 0.45% of total fatty acids, the level of DHA, on the other hand, varies with the mother’s diet from about 0.1%–3.8% of total fatty acids. Unlike breast milk, until relatively recently both term and preterm infant formulas did not contain any n-6 or n-3 PUFAs and it was observed that formula-fed infants have significantly lower levels of DHA in plasma, erythrocytes, and brain cortex compared to breast-fed infants, and lower levels of AA in plasma and erythrocytes (Menon and Dhopeshwarkar 1983). n-3 PUFA supplemented formulas have indeed been shown to be effective in successfully raising infant’s levels of AA and DHA to that of infants who have been fed human milk, within about 10%.

Carlson et al. (1996) were effective in mimicking the levels of AA and DHA in American women’s milk, and when the formula contained 0.1% DHA and 0.43% AA, there were no significant differences in plasma levels of AA and DHA between the breast- and formula-fed groups of infants. Both AA and DHA have to be present in the formula, however, as supplementation with DHA alone has been shown to result in lower levels of AA between 15% and 40% (Auestad et al. 1997). By altering the levels of the longer-chain fatty acids in supplemented formulas and using supplemented formulas (usually containing only LA and ALA) as a reference group, any developmental effects of these manipulations can be investigated.

Carlson et al. (1996) found only a transient benefit of a supplemented formula (0.1% DHA + 0.43% AA) over an unsupplemented formula (LA:ALA = 22:2.2) on visual acuity, which was only present at 2 months but not at 4, 6, 9, and 12 months. In a study using a very similar design and levels of DHA and AA, no advantage was seen in the supplemented group at any testing point (1, 2, 4, 6, 9, and 12 months), although the disparity in results could possibly be due to a different source of fatty acids, i.e., egg phospholipids versus fish oil, respectively (Auestad et al. 2001). Makrides et al. (1995), on the other hand, found that infants fed for 4 months on a supplemented formula (0.36% DHA, 0.58% EPA, 1.52% ALA, and 0.27% y-linolenic acid, n-6) had better transient visual evoked potentials (VEP) at 4 and 7.5 months than the standard 1.6% ALA formula, and the same as the infants fed human milk. Birch et al. (1998) also found that infants given higher levels of DHA in two separate supplemented formulas (0.35% DHA and 0.36% DHA + 0.72% AA) had similar steady-state VEP acuity at 6, 17, and 52 weeks to the infants in the human milk group, and significantly better than the VEP acuity of the standard formula group (LA:ALA = 15:1.5). These results suggest that in terms of visual development, the level of DHA in the diet has to be higher than 0.1% to have a beneficial impact.

This theme is continued as regards the effects of supplemented formulas on cognitive function. Only a handful of studies to date have found a positive impact of added n-PUFAs, and these were with DHA at the levels of 0.35% or 0.36% of total fatty acids (Birch et al. 2000; Birch et al. 2007; Drover et al. 2009). Other studies that have used supplemented formulas where the level of DHA added to the formula was around 0.1% DHA (e.g., Lucas et al. 1999; Makrides et al. 2000; Auestad et al. 2001) have failed to show any differences in cognitive or motor development between infants fed a supplemented formula over the standard one. Interestingly, the level of DHA in American mother’s milk is estimated at around 0.13% DHA, whereas only higher levels of DHA in the formula have been shown to be effective at producing improvements over placebo in these studies.

It is a logical progression to investigate the developmental impact of supplementing the maternal diet with DHA and other n-3 PUFAs (in the absence of a similar n-6 PUFA shortage in the maternal diet). Indeed, the results of a large (N = 11,875) pro-spective epidemiological study, Hibbeln et al. (2007) reported that consumption of less than 340 g of seafood per week was associated with increased risk for suboptimal outcomes for prosocial behavior and fine motor, communication, and social development scores and increased risk for being the lowest quartile for verbal intelligence. Helland et al. (2003) recruited 341 women at 17–19 weeks of their pregnancy and randomly allocated them to a daily regimen of 10 mL of either corn or cod liver oil (1180 mg DHA + 803 mg EPA) until three months after delivery. Plasma levels of DHA were significantly higher in both the infants and the mothers of the cod liver arm compared to the placebo group, demonstrating that maternal dietary supplementation with n-3 PUFAs is reflected in a simultaneous increase in plasma lipid levels of the infant. Fish oil supplementation during pregnancy in this way has been shown to have a positive impact on infant development. The same authors assessed these children at 4 years using the Kaufman Assessment Battery for Children (K-ABC) as an outcome for intelligence and achievement. Infants whose mothers were in the cod liver oil treatment group scored higher on the Mental Processing Composite of the K-ABC, and in a multiple regression model, maternal intake of DHA was the only variable to significantly predict this difference in mental processing at age 4 (Helland et al. 2003); however, these differences disappeared at the 7 year follow-up (Helland et al. 2008). In another randomized double-blind trial study, children whose mothers had been given a fish oil supplement (2.2 g DHA + 1.1 g EPA; N = 33) had better hand–eye coordination at 2.5 years of age than those whose mothers had been given olive oil (N = 39) during pregnancy (Dunstan et al. 2008). There were, however, no significant differences between groups on measures of receptive language or behavior. It is worth noting that maternal supplementation with 2.82 g/day ALA from week 14 of pregnancy to 32 weeks following delivery had no impact on either the infant’s DHA status as measured by plasma lipid levels or on their cognitive function compared to the control group suggesting that the infant requires preformed DHA to meet requirements (de Groot et al. 2004).

Using models of n-3 PUFA deficiency and subsequent repletion, research that has investigated the effects of n-3 PUFAs on behavioral outcomes in animals has demonstrated that brain depletion of n-3 PUFAs occurs in the complete absence of dietary n-3, and is associated with cognitive costs which can be ameliorated once n-3 PUFAs are reintroduced into the diet. Human studies have been far less conclusive. Low n-3 PUFA status is associated with poorer behavioral outcomes, but the evidence provided by intervention studies in the treatment of conditions such as depression, schizophrenia, ADHD, and dementia has been mixed and inconclusive as a whole, although results from a few positive studies have been compelling enough to pursue further research in the area. Overall, the benefit of providing n-3 PUFAs to infants on behavioral outcomes appears to be transient, although the majority of studies have only evaluated the effects of relatively low amounts of DHA, and those providing more than 0.3% DHA in the formula have been more effective. The issue of cogni-tive enhancement via n-3 PUFA supplementation in normally developing children and healthy younger and older adults suggests that supplementation with n-3 PUFAs has little observable effect on behavioral outcomes, even when dietary intake of n-3 PUFAs is low. On the other hand, emerging evidence from neuroimaging studies suggests that supplementation with n-3 PUFAs may be exerting an effect on cerebro-vascular parameters in healthy populations. Future investigations using a variety of imaging techniques to assess the causal relationship between n-3 PUFA intake and brain function in physiological terms are, therefore, warranted.