Deciphering the Role of Docosahexaenoic Acid in Brain Maturation and Pathology with Magnetic Resonance Imaging

Abstract

Animal studies have found that deficits in brain docosahexaenoic acid (DHA, 22:6n-3) accrual during perinatal development leads to transient and enduring abnormalities in brain development and function. Determining the relevance of this evidence to brain disorders in humans has been hampered by an inability to determine antimortem brain DHA levels and limitations associated with a postmortem approach. Accordingly, there is a need for alternate or complementary apsroaches to better understand the role of DHA in cortical function and pathology, and conventional magnetic resonance imaging (MRI) techniques may be ideally suited for this application. A major advantage of neuroimaging is that it permits prospective evaluation of the effects of manipulating DHA status on both clinical and neuroimaging variables. Emerging evidence from MRI studies suggest that greater DHA status is associated with cortical structural and functional integrity, and suggest that reduced DHA status and abnormalities in cortical function observed in psychiatric disorders may be interrelated phenomenon. Preliminary evidence from animal MRI studies support a critical role of DHA in brain maturation. Neuroimaging research in both human and animals therefore holds tremendous promise for developing a better understanding of the role of DHA status in cortical function, as well as for elucidating the impact of DHA deficiency on neuropathological processes implicated in the etiology and progression of neurodevelopmental and psychiatric disorders.

1. Introduction

Mammalian brain tissue is predominantly composed of lipids (60–65% of brain dry weight) which are comprised of saturated, monounsaturated, and polyunsaturated fatty acids. The principle omega-3 polyunsaturated fatty acid in mammalian brain is docosahexaenoic acid (DHA, 22:6n-3), which comprises approximately 10–20% of gray matter, and approximately 2% of white matter, fatty acid composition depending on brain region, age, and habitual dietary omega-3 fatty acid intake [1–6]. Although omega-3 fatty acid precursors of DHA, including α-linolenic acid (ALA, 18:3n-3), eicosapentaenoic acid (EPA< 20:5n-3), and docosapentaenoic acid (22:5n-3), cross the blood-brain barrier, they are rapidly β-oxidized [7,8] and consequently comprise <1% of total brain fatty acid composition [1–6]. Mammals require a dietary source of omega-3 fatty acids to procure and maintain adequate concentrations of DHA in peripheral and central tissues, and healthy adult human exhibit limited or negligible ALA→EPA, ALA→DHA, and EPA→DHA biosynthesis [9]. However, preformed DHA can be obtained directly from the diet, particularly from fatty cold water fish or fish oil supplements [10], and preformed DHA is significantly more effective than ALA for increasing DHA levels in erythrocytes [11], breast milk [12,13], and brain gray matter [14].

Unesterified DHA rapidly diffuses from plasma to brain [15], and at a rate that is equilibrated with brain DHA consumption [16]. DHA preferentially accumulates in gray matter [16,17], and is enriched in synaptic and mitochondrial membranes [18]. DHA is acetylated into the sn-2 position of membrane phospholipids phosphatidylethanolamine and phosphatidylserine [19], and is mobilized preferentially by the calcium-independent phospholipase A2 (iPLA₂) isoform [20]. It has been estimated that approximately 2–8% of rat brain DHA is replaced daily due to metabolism, and has a loss half-life in total rat brain phospholipids of 33 days under steady state ALA intake [21,22]. Dietary ALA insufficiency resulting in deficits in rat brain DHA composition are associated with a reduction in iPLA₂ expression and activity [23], and an increase in the brain DHA half-life [21]. Preliminary estimates of the DHA half-life in human brain phospholipids are 2.5 years [17].

During perinatal rat brain development, cortical DHA concentrations increase sharply in parallel with active periods of neurogenesis, neuroblast migration, differentiation and synaptogenesis [1]. In human brain, DHA accumulates at a rapid rate initiating at approximately the third trimester in utero, and increases to approximately 9% of total cortical fatty acid composition in term-birth infants [24,25]. Infants born preterm exhibit lower postmortem cortical DHA concentrations relative to term infants maintained on the same ALA-fortified formula [25–28]. Non-human primates born preterm similarly exhibit lower postmortem brain DHA concentrations relative to term-born primates [29,30]. During human childhood and adolescence, there is a linear increase in postmortem frontal cortex DHA composition, which stabilizes at ~15% of total cortical fatty acid composition by ~20 years of age [4].

Preclinical studies have provided evidence that brain DHA accrual during perinatal maturation is required for normal neurotrophic factor expression, neurite outgrowth, neurogenesis and migration, neuronal differentiation and dendritic arborization, embryonic cortical plate expansion, nerve growth cone membrane signaling dynamics, and synaptogenesis and plasticity [31–42]. Moreover, early cortical DHA accrual during perinatal development is required for the normal functional maturation of multiple neurotransmitter systems, including dopamine, serotonin, and acetylcholine [43–45]. Behavioral studies have demonstrated that deficits in cortical DHA accrual during perinatal development are associated with enduring impairments in different cognitive tasks [46], and elevated indices of depression and aggression [47]. In addition to the demonstrated neurotrophic effects of DHA, emerging evidence suggests that DHA is protective against neuronal degenerative processes in response to a variety of excitotoxic insults [48–54], and increases resilience of axons and white matter in experimental injury and inflammation models [55–57].

While the importance of cortical DHA accrual in human brain development and function is poorly understood and controversial, a body of evidence suggests that higher maternal DHA status during and following pregnancy is associated with improved infant cognitive development, particularly in the realm of attention [58–65]. Moreover, a growing body of evidence suggests that deficits in attention during childhood frequently precede and predict the subsequent emergence of psychopathology in high-risk populations [66–70]. Importantly, the initial onset of psychiatric disorders, including attention deficit hyperactivity disorder (ADHD), schizophrenia, bipolar disorder, and major depressive disorder (MDD), most frequently occurs during childhood and adolescence [71–73]. It is relevant, therefore, that cross-sectional studies have repeatedly found that patients with ADHD [74–77], schizophrenia [78–80], bipolar disorder [81–83], and MDD [83–89] exhibit peripheral (plasma, erythrocyte) DHA deficits compared with healthy controls. Together, these data suggest that low DHA status may contribute to impaired development of brain circuits that mediate attention.

While the contribution of DHA deficiency to the progressive abnormalities in cortical structure and function observed in psychiatric patients is not known, investigations of the relationship between cortical DHA status and psychopathology have relied on case-control studies of postmortem brain tissue. Some postmortem brain studies have observed significant cortical DHA deficits in patients with psychiatric disorders [90–93] whereas others have not [94–97]. The discrepancy in these findings may be attributable in part to methodological challenges and limitations associated with this approach [98]. Nevertheless, evidence from primary and secondary intervention studies suggest that elevating DHA status through dietary supplementation is efficacious for preventing and/or treating psychopathology in adolescent [99–102] and adult [103–105] patients, and prospective longitudinal studies have found that lower baseline DHA status is a significant predictor of future suicidal attempts in medication-free MDD patients [106] and cytokine-induced MDD [107]. However, it is currently not known if central mechanisms mediate the psychopathogenic effects of low DHA status, and new approaches are required to more definitively elucidate such mechanisms.

2. Clinical MRI studies

Conventional magnetic resonance imaging (MRI) techniques may be well-suited to elucidate the role of DHA in human cortical function and pathology. Modern MRI techniques permit investigation of dynamic changes in cortical structure, chemistry, and functional activity, as well as associated changes in clinical symptoms and DHA status. Because cortical DHA status cannot be determined in living human subjects, peripheral indices (i.e., erythrocyte membrane DHA composition) may serve as a surrogate measure of DHA status. In human subjects, erythrocyte DHA levels provide a valid and reliable index of habitual DHA intake [108–111]. Additionally, non-human primate and human postmortem studies suggest that cortical and erythrocyte DHA levels are positively correlated under steady state dietary conditions [4,5], though erythrocyte DHA levels change more rapidly than brain cortex levels in response to changes in dietary omega-3 fatty acid intake.

Potential neuroimaging techniques available to evaluate the role of DHA status in psychopathology include: (1) structural MRI, which determines cortical and subcortical gray and white matter volumes, (2) diffusion tension imaging (DTI), which determines white matter structural integrity, (3) MRI T₂ relaxometery which determines membrane water content as an index of fluidity, (4) functional magnetic resonance imaging (fMRI), which determines resting and task-elicited changes in cortical activation patterns, (5) proton magnetic resonance spectroscopy (¹H MRS), which determines concentrations of different chemical markers associated with cortical metabolic integrity, and (6) phosphorous magnetic resonance spectroscopy (³¹P MRS), which determines chemical indices of phospholipid membrane turnover. Additionally, positron emission tomography (PET) can determine changes in multiple metabolic processes including cortical fatty acid incorporation and turnover rates and glucose metabolism. While these different imaging techniques have been used extensively in clinical and animal research, only recently have they been employed to investigate the role of DHA status on cortical structural and functional integrity.

The present review will focus on MRI studies investigating the role of DHA in neurodevelopmental and psychiatric disorders, and readers are referred to a separate review in this issue that is focused on the role of DHA in age-related neurodegenerative processes [112]. In the following sections, evidence for abnormalities in cortical structural and functional integrity in psychiatric disorders associated with DHA deficiency is briefly reviewed, and evidence from MRI studies investigating relationships with peripheral indices of DHA status and/or the effects of long-chain omega-3 fatty acid supplementation on MRI outcomes are presented.

2.1. Structural magnetic resonance imaging

Longitudinal and cross-sectional structural MRI studies have begun to characterize gray and white matter maturational patterns in typically developing youth [113]. The childhood and adolescent period is associated with dynamic changes in both regressive (synaptic pruning) and progressive (i.e., myelination) cellular events. Longitudinal structural MRI studies have found that the period between childhood and early adolescence (7–12 years) is associated with a rapid expansion of cortical gray matter density, whereas the period between adolescence (13–18 years) and young adulthood (≥ 8 years) is associated with a progressive loss of cortical gray matter density, which stabilizes in the third decade of life [114]. These age-related changes in cortical volume are sexually dimorphic, peaking later in males than females [115], and are governed by both genetic and environmental factors [116,117]. Postmortem human and non-human primate histological studies suggest that the decrease in cortical gray matter volume during adolescence is attributable in part to reductions in synaptic density rather than neuronal loss [118–121]. Frontal gray matter density loss is associated with reciprocal increases in white matter density in fiber tracts including frontotemporal pathways [122–123], the expansion of which is positively correlated with cognitive development (i.e., performance on working memory tasks)[124].

Although the effect of deficits in prenatal brain DHA accrual on human neuroanatomical trajectories are not known, structural MRI studies have found that children/adolescents born preterm, which is associated with early deficits in cortical DHA accrual, exhibit significant reductions in regional cortical and striatal gray matter volumes, reduced amygdala and hippocampal volumes, reduced corpus callosum and white matter volumes, and larger cerebral ventricles compared with age- and sex-matched term born controls [125]. Importantly, a placebo-controlled structural MRI study found that postnatal DHA supplementation did not significantly alter age-related changes in white matter volume in premature infants [126]. It is also relevant that preterm children are at increased risk for developing ADHD [127–131], and structural MRI studies have found that ADHD children exhibit patterns of cortical gray and white matter volume deficits similar to those observed in preterm children [132].

Emerging evidence from cross-sectional structural MRI studies suggest that patients with MDD exhibit lateral ventricle enlargement and smaller volumes of the basal ganglia, thalamus, hippocampus, frontal lobe, cingulate cortex, and orbitofrontal cortex compared with healthy controls [133]. Bipolar disorder is associated with increased volumes in lateral ventricles, temporal lobe, and putamen [134] as well as progressive frontal white matter pathology [135–138]. Moreover, reductions in amygdala volume have consistently been observed in pediatric and adolescent, but not adult, bipolar patients [139], and chronic lithium exposure is associated with increased hippocampal and amygdala volumes compared with unmedicated patients and healthy controls [134]. A longitudinal structural MRI study further suggests that bipolar disorder is associated with accelerated loss in gray matter volume in subregions of the prefrontal cortex during adolescence compared with typically developing controls [140]. Longitudinal structural MRI studies have also found that schizophrenic patients exhibit greater decreases over time in whole brain volume, whole brain gray matter, frontal gray and white matter, parietal white matter, and temporal white matter volume, as well as greater increases in lateral ventricular volume, compared with healthy controls [141]. Postmortem histological studies suggest that the deficits in gray matter volume observed in patients with psychiatric disorders is attributable in part to lamina-specific reductions in neuronal size and density, dendritic arborization, and/or dendritic spine density rather than neuronal loss [142–149].

Whether these region-specific deficits in gray and white matter volumes observed in psychiatric patients by MRI are associated with lower DHA status, or are preventable or reversible with adequate dietary DHA supplementation, is not known. However, prospective structural MRI is ideally suited to evaluate this relationship. For example, patients with generalized peroxisomal disorders exhibit significant erythrocyte and postmortem cortex DHA deficits [150] and impaired central myelinogenesis [151], and a preliminary structural MRI study found that treatment with DHA ethyl ester (100–600 mg/d) normalized or significantly improved brain white matter volumes in peroxisomal disorder patients [152]. A second preliminary structural MRI study found that greater habitual intake of long-chain omega-3 fatty acids, which are positively correlated with erythrocyte DHA composition [108–111], was associated with larger gray matter volumes in the anterior cingulate cortex, the right hippocampus, and the right amygdala [153]. In a prospective randomized placebo-controlled structural MRI trial, chronic (1 year) treatment with ethyl-EPA was found to slow volume loss in the caudate and thalamus of patients with Huntington's disease [154]. Together, these preliminary MRI findings suggest that DHA status is positively associated with regional gray and white matter volumes.

Neuronal membrane fluidity can also be evaluated using MRI by investigating changes in brain water proton transverse relaxation times (T₂). A preliminary T₂ MRI study found that four week treatment with EPA+DHA decreased whole brain T₂ relaxation time in bipolar patients compared with healthy controls, and was interpreted to reflect an increase cortical membrane fluidity [155]. In a second study, 12-week ethyl-EPA treatment prevented progressive increases in hippocampal T₂ relaxation time observed in placebo-treated first-episode psychosis patients, and reductions in negative symptom severity were associated with smaller increases in T₂ relaxation time [156].

2.2. Diffusion tension imaging (DTI)

As discussed, white matter abnormalities are one of the most consistently reported neuroimaging findings in psychiatric disorders, and preclinical evidence suggests that DHA increases myelin resilience to inflammation and injury. Complimenting MRI determinations of white matter volumes, DTI additionally permits investigation of white matter structural integrity as indexed by fractional anisotropy. A reduction in fractional anisotropy is indicative of loss of white matter integrity, and has been found to be correlated with inter-regional functional connectivity. For example, deficits in functional prefrontal-amygdala connectivity in bipolar patients are correlated with reduced fractional anisotropy in the uncinate fasciculus [157], the principal white matter axonal bundle connecting prefrontal regions with anterior temporal lobe structures [158–160]. While the relationship between DHA status and DTI measures of white matter integrity are not known, a preliminary DTI study (n=12) observed a positive correlation between total plasma polyunsaturated fatty acid concentrations and fractional anisotropy in the uncinate fasciculus white matter tract in medicated psychotic patients [161].

2.3. Functional magnetic resonance imaging

Functional magnetic resonance imaging (fMRI) determines relative changes in blood oxygen level-dependent (BOLD) activity. Increases in BOLD signal are attributable to greater oxygen-dependent synaptic activity and associated increases in local blood volume [162], and reductions in BOLD signal are thought to reflect a suppression of neural activity and shunting of blood from less active to more active regions [163]. fMRI studies can be performed during resting state or during performance of cognitive tasks designed to activate specific brain regions of interest, including the identical pairs continuous performance task. In general, evidence from fMRI studies suggest that psychiatric disorders including MDD and bipolar disorder are associated with greater resting activity in the prefrontal cortex, deficits in task-elicited prefrontal activation, and greater amygdala activation during both resting and active states using different tasks [164–169].

In a dose-ranging controlled fMRI trial, our group investigated the effects of 8-week DHA supplementation on functional cortical activity during performance of sustained attention task (CPT-IP) in healthy male children (8–10 years) [170]. At 8 weeks, erythrocyte membrane DHA composition increased significantly from baseline in subjects receiving low-dose (+47%) and high-dose DHA (+70%), but not placebo (−11%). During sustained attention, both DHA dose groups exhibited significantly greater change from baseline in the activation of dorsolateral prefrontal cortex (DLPFC, BA9), and low-dose and high-dose DHA groups exhibited greater decreases in the occipital cortex and cerebellar cortex, respectively, relative to placebo. Relative to low-dose DHA, high-dose DHA resulted in greater decreases in activation of bilateral cerebellum. Using a less stringent statistical threshold, greater decreases in temporal lobe and cerebellar cortices were observed in the high-dose DHA group compared with placebo. Among all subjects, erythrocyte DHA composition was positively correlated with DLPFC activation at baseline and endpoint. This fMRI study therefore provides proof-of-concept evidence that increasing dietary DHA intake alters functional activity in cortical attention networks in healthy male children. It is relevant, therefore, that mediation-naïve pediatric ADHD patients exhibited reduced prefrontal cortex activation and greater cerebellar cortex activation while performing a sustained attention task [171], a pattern that is opposite to that observed in healthy children receiving DHA supplmentation.

Prior case-control studies have found that MDD patients exhibit DHA deficits in erythrocytes [89] and postmortem prefrontal cortex [91], and EPA+DHA supplementation increases erythrocyte DHA levels and decrease mood symptom severity [100–102]. In view of these findings, our group conducted a prospective 10-week open-label supplementation trial to evaluate the effects of two doses of fish oil (EPA+DHA) (2.4 or 15 g/d) on functional cortical activity patterns in adolescent (10–18 years) MDD patients during performance of a sustained attention task (CPT-IP) [172]. We found that adolescent MDD patients exhibited significantly lower erythrocyte and plasma DHA levels compared with a nested healthy adolescent control group. At baseline, MDD patient erythrocyte DHA composition was positively correlated with functional activation of the prefrontal cortex and anterior cingulate cortex during sustained attention (Fig. 1). Fish oil supplementation significantly increased erythrocyte DHA composition in low- and high-dose groups, and baseline depression symptom severity scores declined significantly in both dose groups. However, we did not observe any significant baseline-endpoint changes in functional activity in either low-dose or high-dose groups, or when both groups were combined. These preliminary fMRI findings suggest that DHA status of adolescent MDD patients is positively associated with functional activity in the prefrontal cortex, and that increasing DHA status reduces depression symptom severity independent of changes in functional cortical activity.

Fig. 1.

Statistical parametric map illustrating the relationship between erythrocyte DHA composition and regional activation patterns during performance of the CPT-IP task at baseline in adolescent MDD patients (n=21). Erythroycte DHA composition is positively correlated with functional activation clustered in the left medial frontal gyrus (BA9, dorsolateral prefrontal cortex) and left anterior cingulate cortex during sustained attention. Significant correlations were defined as an r-value equivalent to p≤0.05 (corrected).

2.4. Proton magnetic resonance spectroscopy (¹H MRS)

Proton magnetic resonance spectroscopy (¹H MRS) is a voxel-based technique that determines cortical concentrations of different compounds including glutamine/glutamate/γ-aminobutyric acid (Glx), myo-inositol (mI), and N-acetyl aspartate (NAA). The Glx peak detected by ¹H MRS is primarily composed of neuronal and astrocyte glutamine, a glutamate precursor, and glutamate. NAA is primarily localized to neurons [173,174] and is positively correlated with mitochondrial metabolism [175,176], and cortical NAA concentrations decrease following excitotoxic injury [177–179]. mI is a product of Gαq-coupled receptor-generated phosphoinositide biosynthesis [180], and is a carbohydrate metabolized from glucose via 1L-myo-inositol 1-phosphate synthase that is predominantly concentrated in astroctyes [173,174]. The mI peak detectable by ¹H MRS likely reflects this astrocyte mI pool.

In general, case-control ¹H MRS studies have observed increased Glx concentrations in the prefrontal cortex of patients with bipolar disorder and schizophrenia, and reduced prefrontal Glx concentrations in patients with MDD [181,182]. ¹H MRS studies have observed decreased NAA concentrations in the prefrontal cortex of patients with bipolar disorder [183–185] and decreased NAA in medial temporal lobe structures of schizophrenic patients [181]. Postmortem studies have observed lower mI concentrations in the prefrontal cortex of patients with unipolar or bipolar depression [186], and some, but not all, ¹H MRS studies have observed reduced mI concentrations in the prefrontal cortex of patients with mood disorders [187–190].

A preliminary placebo-controlled ¹H MRS study found that 12-week supplementation with ethyl-EPA selectively increased NAA concentrations in the anterior cingulate cortex of medicated patients with bipolar disorder [191]. A second placebo-controlled ¹H MRS study found that 12-week ethyl-EPA supplementation increased Glx concentrations in the temporal lobes of first-episode psychotic patients [192]. A third placebo-controlled, dose-ranging, multi-voxel ¹H MRS study, found that 8-week DHA supplementation did not significantly alter any chemical peak including mI, Glx, NAA in the right or left dorsolateral prefrontal cortex (BA9) and anterior cingulate gyrus of healthy male children (8–10 years) [McNamara et al., unpublished data].

2.5. Phosphorous magnetic resonance spectroscopy (³¹P MRS)

³¹P MRS permits central determination of concentrations of phosphorus-containing metabolites, including phospholipid anabolites (i.e., phosphomonoesters, PME) and catabolites (i.e., phosphodiesters, PDE). Reductions in the PME:PDE ratio, secondary to either elevations in PDE or reductions in PME, are thought to reflect a decrease in the synthesis and/or an increase in the breakdown of membrane phospholipids. ³¹P MRS studies have repeatedly observed significant reductions in regional brain PME and concomitant elevations in PDE in the brains of schizophrenic patients [193]. Abnormalities in ³¹P MRS measures of membrane phospholipid metabolism have been observed in medication-naïve first-episode psychosis patients [194–196], and are correlated with both symptom severity [197] and ventricle-to-brain ratio [198]. A meta-analyses of ³¹P MRS studies in patients with bipolar disorder found that PME levels are lower in euthymic patients, and are higher in depressed patients, compared with healthy controls [199].

A preliminary ³¹P MRS study found that cortical PDE levels were correlated with erythrocyte DHA and EPA in healthy subjects [200]. A second study found that erythrocyte total polyunsaturated fatty acids and arachidonic acid (20:4n-6) were positively correlated with bilateral prefrontal cortex PME levels, and linoleic acid (18:2n-6) was positively correlated with PDE levels, in neuroleptic-naïve first-episode psychosis patients [201]. In this study, these fatty acids were not correlated with PME levels in other regions including the basal ganglia, occipital, inferior parietal, or superior temporal cortex. Lastly, a case study found that 6 month treatment with ethyl-EPA resulted in the normalization of brain membrane phospholipid metabolism in a patient with schizophrenia [202].

2.6. Positron emission tomography (PET)

A pioneering PET study using radiolabeled DHA ([1-¹¹C]DHA) investigated incorporation and turnover rates in healthy adult human subjects [17]. This study found that DHA incorporation was greater in gray versus white matter regions, and was positively correlated with regional cerebral blood flow. A second PET study using radiolabeled glucose ([¹⁸F]-fluoro-2-deoxyglucose) evaluated the relationship between plasma DHA composition and resting state cerebral glucose metabolism in adult medication-free MDD patients [203]. Plasma DHA composition was positively correlated with glucose metabolism in the temporoparietal cortex, and negatively correlated with glucose metabolism in prefrontal cortex and anterior cingulate cortex. This study also found that plasma EPA (20:5n-3) levels were not significantly correlated with regional cerebral glucose metabolism. These preliminary data suggest that DHA status is correlated with regional cortical glucose metabolic activity, and that DHA status may have opposing effects on resting activity in prefrontal and temporal cortices.

3. Animal MRI findings

Because it is not possible to directly investigate cortical gray matter DHA composition in living human subjects, analogous MRI techniques are available for rodents so that brain DHA status can be manipulated and quantitated. A rodent 7T Bruker Biospec MRI system is presented in Figure 2A. Moreover, other postmortem brain variables can be directly quantified, including histology (i.e., spine density), gene expression, and neurochemical concentrations to evaluate relationships with MRI measures. Additionally, pharmacological MRI (phMRI) can be used to evaluate dynamic changes in cortical activation patterns following pharmacological challenge [204], and seed-region analysis can be performed to measure functional connectivity between different brain regions [205]. Rodent MRI is therefore well-suited for extending and interpreting clinical MRI findings.

Fig. 2.

The 7T Bruker Biospec rodent imaging system (A), localization of a ¹H MRS voxel in rat bilateral medial prefrontal cortex in the coronal view (+3.7 cm anterior of Bregma)(B), and a representative ¹H MRS spectrum from a control rat medial prefrontal cortex (C). Cr, creatine; Glx, glutamate+glutamine; mI, myo-inositol; Cho, choline; NAA, N-acetyl aspartate.

To date there have been few studies investigating the effects of brain DHA deficits or enrichment on MRI outcomes. A structural MRI study did not find alterations in brain gray or white volumes in aged (>15 months) omega-3 fatty acid-deficient rats relative to aged controls [206]. In a structural MRI study, our group recently found that young adult rats subjected to deficits in cortical DHA accrual during postnatal development exhibited elevations in volume in prefrontal cortex, nucleus accumbens, and amygdala in young adulthood compared with controls [207]. In an ¹H MRS study, our group investigated the effects of perinatal deficits in DHA accrual on Glx, mI, and NAA concentrations in rat medial prefrontal cortex. The position of the voxel in the rat medial prefrontal cortex and a representative ¹H MRS spectrum is presented in Figure 2. We found that perinatal, but not postnatal, deficits in brain DHA accrual were associated with reductions in baseline concentrations of mI, but not Glx or NAA, in the medial prefrontal cortex [208]. Additionally, acute treatment with SKF83959, a selective agonist at dopamine D₁ phosphoinositide-coupled receptors, increased mI concentrations in the perinatal deficiency group but not in controls [208]. To date, there have been no animal ³¹P-MRS, phMRI, or DTI studies conducted to investigate membrane turnover, functional activity and connectivity, or white matter integrity, respectively, in DHA-deficient or DHA-enriched rat brain.

4. Discussion

Emerging evidence from both human and animal MRI studies are beginning to develop a clearer understanding of the role of DHA in cortical structure and function, and may provide critical insight into the contribution of DHA deficiency to the progressive neuropathological brain changes implicated in psychiatric disorders. An advantage of clinical MRI is that it permits prospective and concomitant evaluation of the effects of manipulating DHA status on both symptom severity and different neuroimaging variables to investigate potential central mediating mechanisms. A limitation associated with clinical MRI studies includes reliance on peripheral measures of DHA status which may not accurately reflect brain DHA levels [209]. However, emerging evidence suggests that plasma and/or erythrocyte DHA levels are correlated with functional cortical activity by fMRI [170], membrane phospholipid turnover by ³¹P MRS [200], and resting cortical glucose metabolism by PET [203]. Although rodent MRI studies permit systematic manipulation of brain DHA composition, rodent brain DHA deficiency models frequently produce large reductions in brain DHA levels which may not be clinically relevant. Another limitation of rodent MRI studies is the requirement for general anesthesia to obviate motion artifacts, and different general anesthetics have been found to alter blood blow dynamics [204] and neuronal morphology [210]. Nevertheless, combining evidence from both clinical and animal MRI approaches is anticipated to accelerate our understanding of the role of DHA in cortical structure and function.

Regarding future MRI studies, evaluation of the relationship between DHA status and functional brain maturation requires a prospective longitudinal study design. By analogy, prior prospective longitudinal MRI studies have found robust effects of psychostimulant medications on cortical developmetal trajectories in ADHD patients [211], and that semi-chronic lithium treatment increased prefrontal cortical volumes in bipolar patients responding to treatment [212]. Using this approach, it will be of considerable interest to determine whether increasing DHA status can normalize abnormlities in cortical gray and white matter volumes in psychiatric patients, and whether increasing DHA status can prevent or forestall progression of neuropathology in high-risk subjects. In an ongoing placebo-controlled trial (NCT00917501) we are evaluating whether 12-week EPA+DHA (fish oil) supplementation influences functional connectivity by fMRI and chemical indices of metabolic integrity by ¹H MRS in medication-free adolescents at ultra high-risk for developing mania (i.e., they have a biological parent with bipolar disorder and are diagnosed with DSM-IV MDD).

An issue for future omega-3 fatty acid intervention trials is brain bioavailability. Although preformed DHA is effective for increasing cortical gray matter DHA levels [14], unesterified DHA, but not phospholipid-esterified DHA, diffuses from plasma to brain more efficiency [16]. By analogy, peripheral bioavailability of EPA+DHA from re-esterified triglycerides was found to have greater incorporation into peripheral membranes compared with natural fish oil [213]. Moreover, our preliminary fMRI data suggest that supplementation with triglyceride DHA [170], but not natural fish oil [172], increases functional prefrontal cortical activity. Future fMRI studies are therefore warranted to determine which DHA carriers are most effective for altering functional cortical activity. Moreover, EPA is rapidly oxidized following entry into brain [7], plasma EPA levels are not correlated with resting cortical glucose metabolism in MDD patients [199], and EPA→DHA biosynthesis is negligible in human subjects [9]. However, MRI studies have observed effects of ethyl-EPA supplementation on different aspects of cortical metabolic function [154,156,191,192,202], raising the possibility that peripheral actions of long-chain omega-3 fatty acids may also contribute to central changes observed by MRI.

In view of the clinical observation that patients with psychiatric disorders exhibit deficits in frontal white matter volume by structural MRI and reduced frontal white matter integrity by DTI, and preclinical evidence that increasing brain DHA status increases resilience of white matter to experimental injury and inflammation, it will be of considerable interest to determine in future studies whether increasing DHA status can reverse deficits in frontal white matter integrity in patients with psychiatric disorders. Furthermore, preclinical evidence that increasing brain DHA status increases dendritic spine density [41], and it will be of interest to determine whether DHA supplementation can alter developmental trajectories in human gray matter volume and associated cognitive function in typically developing youth, as well as slow accelerated gray matter atrophy observed in patients with psychiatric disorders.

These preliminary observations support a role for MRI research to develop a clearer understanding of the role of DHA in cortical functional maturation as well as for elucidating the contribution of DHA deficiency to neuropathological processes implicated in the pathoetiology of psychiatric disorders. While human and animal MRI techniques both have unique limitations, combining these approaches is anticipated to accelerate our understanding of the role of DHA in cortical structure and function as well as guide the optimization of treatment and prevention strategies.