152 (2007) 1–10www.elsevier.com/locate/psychres

Psychiatry Research

Serum ω-3 fatty acids are associated with variation in mood,personality and behavior in hypercholesterolemic

community volunteers

Sarah M. Conklin a,⁎, Jennifer I. Harris b, Stephen B. Manuck c, Jeffrey K. Yao d,Joseph R. Hibbeln e, Matthew F. Muldoon f

a Cardiovascular Behavioral Medicine Postdoctoral Training Program, Department of Psychiatry,University of Pittsburgh, Pittsburgh, PA, United States

b Department of Psychiatry, Brown University, Providence, RI, United Statesc Department of Psychology, University of Pittsburgh, Pittsburgh, PA, United States

d VA Pittsburgh Healthcare System and Departments of Psychiatry and Pharmaceutical Sciences,University of Pittsburgh, Pittsburgh, PA, United States

e National Institute on Alcohol Abuse and Alcoholism, Bethesda, MD, United Statesf School of Medicine, University of Pittsburgh, Pittsburgh PA, United States

Received 9 July 2006; received in revised form 12 September 2006; accepted 1 October 2006

Abstract

Low dietary intake ofω-3 polyunsaturated fatty acids has been linked to several features of psychiatric symptomatology, includingdepression, disorders of impulse control, and hostility. Preliminary intervention trials of ω-3 fatty acid supplementation for clinicaldepression and other disorders have reported benefit. However, few studies have investigated the relationships between these fattyacids and normative variability in mood, behavior and personality. Participants were 105 hypercholesterolemic, but otherwise healthy,non-smoking adults. Fasting serum α-linolenic (α-LNA), eicosapentaenoic (EPA) and docosahexaenoic acid (DHA) were assayedwith gas chromatography. Participants completed the Beck Depression Inventory (BDI), the NEO Five Factor Personality Inventory(NEO-FFI) and the Barratt Impulsiveness Scale (BIS). In multivariate analyses, higher levels of the long chain ω-3 PUFAs, EPA andDHA, were associated with significantly reduced odds of scoring≥10 on the BDI. Similarly, DHA and EPA covaried inversely withNEO-Neuroticism scores, whereas DHAwas positively associated with NEO-Agreeableness. On the BIS, DHAwas inversely relatedto cognitive impulsivity and α-LNAwas inversely related to motor and total impulsivity. These findings suggest that ω-3 fatty acidstatus is associated with variability in affect regulation, personality and impulse control.© 2006 Elsevier Ireland Ltd. All rights reserved.

Keywords: Polyunsaturated fatty acids (PUFA); Depression; Impulsivity; Personality

⁎ Corresponding author. University of Pittsburgh, Department ofPsychiatry, 506 Old Engineering Hall, 4012 O'Hara Street, Pittsburgh,PA 15260, United States. Tel.: +1 412 624 8828; fax: +1 412 624 9108.

E-mail address: conklinsm@upmc.edu (S.M. Conklin).

0165-1781/$ - see front matter © 2006 Elsevier Ireland Ltd. All rights resedoi:10.1016/j.psychres.2006.10.006

1. Introduction

The ω-3 polyunsaturated fatty acids (PUFAs) mustbe obtained in the diet and a deficiency has been linkedto several major neurobehavioral disorders. As reviewedelsewhere, a relative deficiency in dietary ω-3 PUFAs

rved.

2 S.M. Conklin et al. / Psychiatry Research 152 (2007) 1–10

has been implicated in major depressive disorder(Hibbeln and Salem, 1995; Freeman, 2000; Logan,2003; Peet and Stokes, 2005), attention deficit hyper-activity disorder (Richardson and Puri, 2002), impulsecontrol problems, including aggressive behavior (Hal-lahan and Garland, 2004), and schizophrenia (Skosnikand Yao, 2003). In animals, dietary ω-3 PUFAdeficiency alters brain chemistry, development andneurotransmission (Das, 2003; De la Pressa Owensand Innis, 2003). Related literature suggests that a highintake of ω-6 PUFAs (principally arachidonic acid[AA]), relative to dietary ω-3 PUFA consumption,contributes to these same psychological disorders(Clandinin et al., 1992).

These associations are plausible reflections of theroles played by PUFAs in the brain. The principal long-chain, ω-3 polyunsaturated fatty acids – eicosapentae-noic and docosahexaenoic acids (EPA and DHA) – mayeither be consumed by eating seafood or synthesizedfrom α-linolenic acid, a medium chain ω-3 PUFA,found in certain vegetables. DHA is concentrated in thehuman brain and constitutes more than 30% of fattyacids in grey matter (Salem et al., 2001). These fattyacids occupy the sn-2 position of neural membranephospholipids and, in that position, regulate iontransport and G-protein and receptor function (Clandi-nin et al., 1992; Hulbert et al., 2005). Theω-6 fatty acidsare also released from membranes to influence thebody's responses to inflammatory and oxidative stres-ses. Moreover, dietary ω-3 PUFA intake alters theexpression of many genes, including several that areimportant for synaptic plasticity and learning (Kitajka etal., 2004). PUFAs of the ω-6 family may also occupythe sn-2 position and do so to a greater degree whendietary ω-3 consumption is low. This ‘replacement’ isbelieved to disrupt an optimal ω-6:ω-3 membrane ratioand increase vulnerability to pathogenic inflammatoryprocesses, (Salem et al., 2001; Calder and Grimble,2002; Song et al., 2003; Yao and Reddy, 2003).

Population studies have found that low dietary intakeof ω-3 and high ω-6:ω-3 ratios is associated with highrates of depression (Hibbeln, 1998, 2002; Tanskanenet al., 2001a), whereas populations with higher fishconsumption have lower rates of postpartum depression(Hibbeln, 2002) and suicide (Tanskanen et al., 2001b).Other observational evidence links low serum ω-3 and/or high ω-6 levels with depressive symptomatology(Adams et al., 1996; Mamalakis et al., 2002 andTimonen et al., 2004), major depressive disorder(Edwards et al., 1998; Frasure-Smith et al., 2004;Maes et al., 1999; Peet et al., 1998), and other forms ofnegative affect, including hostility (Iribarren et al.,

2004) and anxiety (Green et al., 2006). In some but notall preliminary clinical trials, supplementation with thelong-chain ω-3 PUFAs has been shown to improvemood in patients with major depression and bipolardisorder (Nemets et al., 2002; Peet and Horrobin, 2002;Stoll et al., 1999; Su et al., 2003; Marangell et al., 2003;Ness et al., 2003).

EPA and DHA also have been implicated inimpulsive and aggressive behaviors. Low serum ω-3PUFAs have been found significant predictors ofaggression in psychiatric inpatients (Buydens-Brancheyet al., 2003b), relapse in cocaine addicts (Buydens-Branchey et al., 2003a), and frequency of externalizingbehavior problems in boys (Stevens et al., 1996). Anumber of small randomized controlled trials supportthese associations. A significant attenuation of aggres-sive behavior was observed among women withBorderline Personality Disorder (Zanarini and Franken-burg, 2003) followingω-3 supplementation alone and aspart of multi-nutrient intervention among prisoners(Gesch et al., 2002).

It is not known whether the associations betweenserum ω-3 fatty acids, negative affect and impulsivityobserved in clinical samples extend to normal variationin these psychological and behavioral characteristics,although a small randomized trial of fish oil supple-mentation showed impressive improvements of moodand reaction time in healthy participants (Fontani et al.,2005). Accordingly, the purpose of the present inves-tigation was to determine if serum ω-3 fatty acids arecorrelated with differences in depressive symptomatol-ogy, personality, and impulsivity in a non-psychiatriccommunity sample. Fatty acid status was determined infasting serum samples. PUFAs are supplied to the brainvia blood, and therefore serum concentration provides ameasure of nutrient delivery to the brain. We hypoth-esized that low serum concentrations of the ω-3 fattyacids – α-LNA, EPA and DHA – and elevated ω-6:ω-3ratios would be associated with negative affectivity andimpulsivity in community adult volunteers.

2. Methods

2.1. Participants

This investigation was a substudy of the Cholesteroland Risk Evaluation project, phase 2 (CARE-2), a studyof relationships between lipids and behavior (Muldoonet al., 2004). Participants were recruited through massmailings and advertisements. A total of 308 volunteersmet criteria for inclusion in CARE-2. Participants werestratified by sex, race, and median age, and randomly

Table 1Characteristics of participants (N=105)

Fatty acid concentrations Median Inter-quartile range(25%–75%)

Serum total fatty acids (μg/ml) 3224.8 2730.7–3700.4ω-3 fatty acids (μg/ml)DHA 46.4 36.9–57.9α-LNA 19.5 13.8–27.5EPA 18.1 12.7–24.1

ω-6 fatty acids (μg/ml)LA 971.2 843.3–1120.6AA 255.1 225.1–296.7

ω-6:ω-3 ratiosLA:α-LNA 50.5 40.9–60.9AA:EPA 14.1 10.9–18.3AA:DHA 5.5 4.5–7.2

Psychological measure Mean (SD) Range (minimum–maximum)

Beck Depression Inventory 5 (5) 0–27Barratt Impulsiveness Scale

Total 33 (11) 9–67Motor 11 (5) 2–28Cognitive 10 (4) 2–20Non-planning 13 (6) 1–30

NEO-Five Factor InventoryNeuroticism 14 (8) 0–39Extraversion 29 (7) 3–42Openness 25 (7) 10–40Agreeableness 35 (6) 17–47Conscientiousness 36 (6) 19–47

3S.M. Conklin et al. / Psychiatry Research 152 (2007) 1–10

assigned to one of three groups (placebo, 10 mg or40 mg simvastatin) for 6 months. Serum fatty acids weredetermined on the first 106 consecutive participantsassigned to either placebo or 40 mg simvastatin whowere non-smokers and reported not taking fish-oilsupplements. This report constitutes cross-sectionalanalyses of baseline data. Participants were between35 and 70 years of age, and all had untreatedhypercholesterolemia, defined as an average low-density lipoprotein (LDL) cholesterol between 160 and220 mg/dl. Exclusion criteria included the use of anypharmacotherapy for lipid or psychiatric disorders, and ahistory of coronary heart disease, diabetes, schizophre-nia, neurological disorders, hepatic or renal insufficien-cy, marked hypertriglyceridemia (N360 mg/dl), cancerand untreated thyroid disorders. Hypertensive indivi-duals were considered eligible if their antihypertensivetreatment was stable for the previous 3 months. Theprotocol was approved by the University of PittsburghInstitutional Review Board and subjects gave informedconsent.

Frozen serum samples were analyzed for fatty acidcontent and composition. None of the participants in thissubstudy smoked, took fish oil supplements, consumedover two alcoholic drinks per day or had gastrointestinaldisorders that could affect absorption. One participantdid not complete the psychological questionnaires, sothat 105 participants were included in the final analyses.

2.2. Serum fatty acid quantification

Serum fatty acid composition was quantified as aconcentration (μg/ml) and as a percentage of the totalfatty acid pool by gas chromatography as describedpreviously (Harris et al., 2004). Intra-assay variations ofb1% for repeated chromatographic assays and b5% fortotal serum sample preparation were typical. The currentreport describes analyses of baseline levels of three ω-3fatty acids — 18:3n-3, α-linolenic acid (α-LNA), 20:5ω-3, eicosapentaenoic (EPA) and 22:6 ω-3, docosahex-aenoic acid, DHA. In addition, three ratios wereexamined — AA:EPA, AA:DHA, and LA:α-LNA(20:4 ω-6, arachidonic acid (AA), 18:2 ω-6, linoleicacid (LA)).

2.3. Self-report measures

Three self-report measures of mood and personalitywere administered: the Beck Depression Inventory (BDIBeck et al., 1988); the Barratt Impulsiveness Scale (BIS;Barratt, 1985); and the NEO Five Factor Inventory(NEO-FFI; Costa and McCrae, 1992).

2.3.1. Beck Depression InventoryThe BDI is a 21-item self-report measure designed to

assess current depressive symptomatology. By con-vention, total scores are categorized as follows: 0–9,none to minimal depression; 10–18, mild-to-moderate;19–29, moderate-to-severe depression; N29, and severedepression (Beck et al., 1988; Tanskanen et al., 2001a).A meta-analysis for internal consistency yielded a meanalpha of 0.81 for non-psychiatric samples (Beck et al.,1988).

2.3.2. NEO Five-Factor InventoryThe NEO-FFI is an abbreviated 60-item version of

the revised NEO Personality Inventory (NEO-PI; Costaand McCrae, 1992). The instrument assesses fivedimensions of adult personality (neuroticism, extraver-sion, openness to experience, agreeableness and con-scientiousness) commonly observed on factor-analyticreduction of lexically-derived trait descriptors. All itemsare scored on a 5-point likert scale. The NEO-FFI hasgood internal consistency (0.74 to 0.89) and retestreliability (0.83–0.87) (Costa and McCrae, 1992).

Fig. 1. Odds ratios and 95% confidence intervals for depressivesymptomatology (BDI≥10) for a one SD increase in fatty acidmeasures. The figure illustrates a significant reduction in the odds forscoring ≥10 on the BDI with each SD increase in EPA expressed as apercentage of the total serum fatty acids (OR=0.051, 95% CI=0.27–0.97) and for a 1 SD increase in DHA expressed as a concentration(OR=0.46, 95% CI=0.24–0.88) or as a percentage of the total fattyacids (OR=0.42, 95% CI=0.21–0.84.

4 S.M. Conklin et al. / Psychiatry Research 152 (2007) 1–10

2.3.3. Barratt Impulsiveness ScaleThe BIS is a 30-item questionnaire designed to assess

control of thoughts and behavior (e.g. acts withoutthinking, decides “on the spur of the moment,” does notplan ahead) each scored on a four point likert scale(Barratt, 1985; 1994). A total score is obtained by

Table 2Correlations between serum fatty acids1 and measures of mood, personality

Measure LNA EPA

μg/ml % of total μg/ml % of total

BDI2

Total −0.01 0.00 −0.19 # −0.20⁎

NEO-FFI3

N 0.02 −0.04 −0.21⁎ −0.28⁎E −0.07 0.09 0.03 0.06O 0.06 0.10 0.00 0.00C 0.15 0.17 0.06 0.02A 0.09 0.10 0.12 0.10

Barratt Impulsiveness Scale4

Motor −0.23⁎ −0.32⁎⁎ −0.04 −0.01Cog. −0.10 −0.16 −0.24⁎ −0.28⁎Non-plan. −0.12 −0.11 −0.09 −0.04Total −0.20⁎ −0.26⁎⁎ −0.15 −0.13

Note: Point-biserial and Pearson correlation coefficients, n=105; ⁎P≤0.05;1Serum fatty acids are expressed as both a concentration and as a percentage othe point-biserial-correlation coefficients based on the dichotomized group[≥10; n=17]). 3NEO Five-Factor Inventory. N, neuroticism; E, extraverCognitive Impulsiveness Scale; Non-plan.=Non-planning Impulsiveness Sc

summing all items, and three subscales derived by factoranalysis are labeled ‘motor,’ ‘cognitive’ and ‘non-planning’ impulsiveness (Patton et al., 1995). The BIShas high internal consistency (alpha coefficients 0.79–0.83) and, although retest reliability of the instrumenthas not been reported, we have observed highreproducibility of scale scores over an average retestinterval of 3 years among 96 adult men (reliabilitycoefficient 0.85; retest reliabilities were 0.77 for themotor scale, 0.73 for cognitive impulsiveness and 0.77for non-planning) (Manuck et al., 2000).

2.4. Statistical analyses

All fatty acid measures were log-transformed tonormalize data distributions. Because most subjects hadlow scores on the BDI, the conventional cut-off value of10 or greater (Beck et al., 1988) was used to dichotomizesubjects into two groups: no or minimal depressivesymptoms (n=88, range 0–9) and mild or greaterdepressive symptomatology (n=17, range 10–27).

Simple bivariate associations between fatty acids andpsychological test scores were first examined usingpoint-biserial correlations for the BDI and Pearsonproduct-moment correlations for the BIS and NEO-FFI.Significant associations (P≤0.05) were further exam-ined using multivariate regression analyses controllingfor gender, race, and age. BDI scores were analyzed by

and impulsivity

DHA LA:LNA AA:EPA AA:DHA

μg/ml % of total

−0.22⁎ −0.22⁎ 0.00 0.16 0.17

−0.25⁎ −0.32⁎⁎ 0.11 0.20⁎ 0.22⁎

0.10 0.14 0.05 0.06 0.02−0.07 −0.07 −0.05 −0.05 0.000.10 0.04 −0.11 −0.10 −0.150.24⁎ 0.20⁎ −0.16 −0.09 −0.19

−0.05 0.03 0.28⁎⁎ 0.05 0.05−0.20⁎ −0.22⁎ 0.19 0.28⁎⁎ 0.24⁎

−0.13 −0.06 0.05 0.06 0.09−0.17 −0.12 0.22⁎ 0.16 0.16

⁎⁎P≤0.005; # P=0.058.f the total fatty acids. 2BDI=Beck Depression Inventory; row containsmembership (0= low BDI group [b10; n=88] or 1=high BDI groupsion; O, openness; C, conscientiousness; A, agreeableness. 4Cog.=ale.

5S.M. Conklin et al. / Psychiatry Research 152 (2007) 1–10

logistic regression and the odds of having at least milddepressive symptomatology was estimated for a onestandard deviation increase in each fatty acid measure.In linear regression analyses of BIS and NEO-FFIscores, gender, race and age were entered in step 1 andeach fatty acid measure was entered individually in step2. The proportion of variance in NEO or BIS scoresaccounted for by covariates (R2) was obtained in step 1,and the additional variance accounted for by each fattyacid (ΔR2) was obtained in step 2.

3. Results

The study sample, which was 49% male and 83%Caucasian, had a mean age of 54±9 years and meanbody mass index of 29±5 kg/m2. The mean serum totaland low-density lipoprotein cholesterol concentrationswere 264±24 and 181±16 mg/dl, respectively. De-scriptive statistics for serum fatty acid concentrations,and psychological test scores are provided in Table 1.

First examined were potential differences in fattyacid and psychological measures as a function ofgender, race and age. Women scored higher than menon NEO Agreeableness (Pb0.05), whites had a lower

Table 3Multiple regression analyses assessing the relationship between fatty acids a

Behavioral measurements Serum concentrations (μg/ml)

α-LNA EPA

β ΔR2 P β

NEO-FFINeuroticism −0.001 0.000 ns −0Agreeableness 0.101 0.010 ns 0

Barratt Impulsiveness ScaleCognitive −0.102 0.010 ns −0Motor −0.273 0.071 0.006 −0Total −0.213 0.044 0.034 −0

Behavioral measurements Percentage of total fatty acid pool

α-LNA EPA

β ΔR2 P β

NEO-FFINeuroticism −0.046 0.002 ns −0Agreeableness 0.115 0.013 ns 0

Barratt Impulsiveness ScaleCognitive −0.163 0.026 ns −0Motor −0.343 0.115 <0.001 −0Total −0.263 0.068 0.008 −0

Note: All PUFA values were log transformed; β=standardized value; Cova(1=male, 2=female) entered simultaneously on the first step of the regressionfatty acid.

mean AA: DHA ratio than non-whites (Pb0.05), and aweak negative correlation was found between age andthe AA:DHA ratio (r=−0.182, P=0.062). In view ofthese associations, gender, race and age were includedas covariates in all multivariate analyses.

Simple correlations between serum fatty acids andscores on the psychological instruments are presented inTable 2. Serum levels of both EPA and DHA variedinversely with BDI scores, indicating that low concen-trations of long-chain ω-3 fatty acids were associatedwith greater depressive symptomatology. With respectto NEO-FFI personality traits, both EPA and DHAcorrelated negatively with Neuroticism. Consistent withthese latter findings, ratios of AA: EPA and AA: DHAcorrelated positively with Neuroticism. Serum DHAconcentration correlated positively with Agreeableness.On the BIS, α-LNA, the 18-carbon essential ω-3 fattyacid, correlated negatively with BIS Motor and Totalimpulsivity scores. A negative correlation was alsofound between both serum EPA and DHA and the BISCognitive impulsivity score. Ratio's of AA:EPA andAA:DHA correlated positively with the Cognitivescale, whereas, the LA:α-LNA ratio was positivelyassociated with Motor and Total scores. Thus, greater

nd measures of impulsivity and personality

DHA

ΔR2 P β ΔR2 P

.212 0.044 0.033 −0.246 0.060 0.013

.140 0.019 ns 0.217 0.047 0.025

.236 0.054 0.018 −0.214 0.045 0.031

.055 0.003 ns −0.055 0.003 ns

.150 0.022 ns −0.172 0.029 ns

DHA

ΔR2 P β ΔR2 P

.282 0.076 0.005 −0.318 0.095 0.002

.133 0.017 ns 0.199 0.037 0.046

.279 0.075 0.005 −0.240 0.054 0.018

.003 0.000 ns −0.022 0.001 ns

.124 0.015 ns −0.121 0.014 ns

riates were age (in years), race (1=white, 2=non-white) and genderequation;ΔR2 reports the proportion of variance accounted for by each

6 S.M. Conklin et al. / Psychiatry Research 152 (2007) 1–10

ω-6:ω-3 fatty acid ratios were associated with greaterimpulsivity.

Significant bivariate associations were re-examinedusing logistic or linear regression, controlling forgender, race and age. A total of 23 bivariate associationswere significant at Pb0.05, and 1 approached signifi-cance with P=0.058; as a result, 24 associations wereexamined further with multivariate adjustment. Asdepicted in Fig. 1, a one SD increase in DHA wasassociated with a reduction in the odds of depressivesymptomatology (BDI score≥ 10) by over 50%(P=0.018 and 0.015 for serum concentration andpercentage of total fatty acids, respectively). Similar,but somewhat weaker findings were noted for serumEPA (P=0.063 and 0.038 for serum concentration andpercentage of total fatty acids, respectively). Notably,the low and high BDI groups did not differ onphysiological parameters influenced by dietary intake,including serum triglycerides and body mass index(BMI).

Results of multiple linear regression analyses forNEO-FFI and BIS scores are displayed in Table 3.Analyses of the fatty acids as a concentration (μg/ml)and as a percentage of the total pool are displayed in theupper and lower panels, respectively. Results matchedthose of the bivariate analyses and generally indicatedthat lower concentrations and percentages of the ω-3fatty acids were related to higher neuroticism andimpulsivity scores. In addition, there was a positiveassociation between DHA and agreeableness. Table 4reports the results of multiple linear regression analysesof BIS and NEO-FFI scores using theω-6:ω-3 fatty acidratios as predictor variables. Again as found in thebivariate analyses, higher ω-6:ω-3 ratios in severalinstances predicted higher neuroticism and impulsivityscores, and agreeableness was marginally predicted by alow AA:DHA ratio (P=0.059).

Table 4Multiple regression analyses assessing the relationship between serum fatty

Behavioral Measurements LA:α-LNA AA

β ΔR2 P β

NEO-FFINeuroticism 0.115 0.013 ns 0Agreeableness −0.173 0.029 ns −0

Barratt Impulsiveness ScaleCognitive 0.186 0.034 ns 0Motor 0.308 0.091 0.002 0Total 0.218 0.048 0.026 0

Note: All PUFA values were log-transformed; β=standardized value; Cova(1=male, 2=female) entered simultaneously on the first step of the regressi

4. Discussion

The present study reports associations between serumω-3 PUFAs and normative variability of negative affectand impulsivity. In bivariate analyses and after multi-variate adjustment for age, sex and race, neuroticism,depressive symptomatology and cognitive impulsivitywere all associated inversely with both EPA and DHA,whether expressed as a serum concentration or as apercentage of the total fatty acid pool. Additionally, NEOagreeableness correlated positively with DHA. Ouranalyses indicate that high ω-6:ω-3 ratios were associ-ated with the same psychological characteristics thatcorrelated negatively with serum ω-3 concentrations. Insum, these findings suggest that serum ω-3 fatty acidsare a significant correlate of trait variation in affectregulation, personality and impulse control, as seen in anon-psychiatric and healthy (albeit hypercholesterol-emic), community sample of adult men and women.

Although the associations described here are novelwith regard to the sample characteristics, these findingsare consistent with a number of previous investigationsin psychiatric populations. For example, our data accordwith previously reported associations between lowerblood ω-3 PUFAs in clinically depressed patients (Ed-wards et al., 1998; Maes et al., 1999) and in patients withbipolar disorder (Chiu et al., 2003; Stoll et al., 1999;Zanarini and Frankenburg, 2003). Similarly, low serumlevels and/or dietary intake of ω-3 PUFA have beenimplicated in psychopathologies for which problems ofimpulse control figure prominently, including substancedependence (Buydens-Branchey et al., 2003a), border-line personality disorder (Stoll et al., 1999; Zanarini andFrankenburg, 2003), suicide (Huan et al., 2004) andADHD (Richardson and Puri, 2002; Richardson andMontgomery, 2005). Here, we extend these findings byshowing low serum ω-3 PUFAs associated with higher

acid ratios and measures of impulsivity and personality

:EPA AA:DHA

ΔR2 P β ΔR2 P

.221 0.045 0.031 0.228 0.050 0.023

.130 0.016 ns −0.186 0.033 ns

.271 0.073 0.004 0.251 0.060 0.012

.079 0.006 ns 0.078 0.006 ns

.161 0.024 ns 0.165 0.026 ns

riates were age (in years), race (1=white, 2=non-white) and genderon.

7S.M. Conklin et al. / Psychiatry Research 152 (2007) 1–10

scores on the Barratt Impulsiveness Scale in a non-patient sample.

The few previous studies that have examined relation-ships between ω-3 fatty acids and psychologicalcharacteristics and behaviors in non-patient sampleshave generally relied upon ω-3 intake estimated fromdietary assessments (Daviglus et al., 1997; Iribarren et al.,2004; Tanskanen et al., 2001a,b). Serum ω-3 PUFAsnecessarily reflect dietary intake, yet serum concentra-tions of the essential and long-chainω-3 PUFAs correlateonly modestly with ω-3 intakes from dietary assessment(Hjartaker et al., 1997). This is perhaps due to biologicalvariation between individuals (e.g. absorption, metabolicdifferences) and to the recall bias and errors inherent indietary questionnaires and interviews. In the currentinvestigation, fasting serum samples were assayed forfatty acid composition. This approach may circumventthe limited accuracy of self-reported measures of dietaryintake, which may not account for the competitive intakesof ω-6 fatty acids, and also provides an estimate of ω-3PUFA delivery to the brain.

The current investigation is cross-sectional and,therefore, cannot ascertain causal association. One pos-sibility is that mood, impulsivity and personality deter-mine and/or influence a person's dietary choices. Thisnotion is supported by evidence that certain character-istics of personality, such as neuroticism, are associatedwith poor health behaviors (Adler and Matthews, 1994;Russo et al., 1997).

Alternatively, dietary ω-3 intake could affect mood,personality and behavior (Fontani et al., 2005). Anumber of small controlled trails have reported improve-ments in mood and reductions in impulsive and ag-gressive behaviors following increases in dietary ω-3(Hamazaki et al., 1999, 2002, 2005; Itomura et al., 2005).This alternative explanation is supported by a number ofplausible biological mechanisms. Dietary PUFAs arerapidly delivered to the brain via blood (Arab, 2003;Clandinin et al., 1992; Rapoport et al., 2001). Once in thebrain, PUFAs incorporate into neural membranes byanchoring within the phospholipids and influence thephysical properties of the membrane sometimes referredto as fluidity. Although DHA (22:6n-3) differs in onlyone double bond from the ω-6 replacement 22:5n-6, thisincreased flexibility in the methyl end allows for a moreeven distribution of chain densities along the bilayer,which affects several important membrane-dependentprocesses such as ion flux, G-protein activity, receptorand transmitter function (Salem et al., 2001; Eldho et al.,2003). In laboratory animals, PUFA-deficient diets fed inearly development reduces frontal cortex monoaminer-gic concentrations of dopamine and serotonin as well as

their degradative metabolites, 3,4-dihydroxy-phenyla-cetic acid (DOPAC), homovanillic acid (HVA) and 5-hydroxyindolacetic acid (5-HIAA), respectively (De laPressa Owens and Innis, 2003). PUFA intake also affectsbrain development in humans. Babies fed formula defi-cient in long-chain PUFAs have lower brain DHA andpoorer cognitive functioning when compared to breast-fed infants (Anderson et al., 1999; Farquharson et al.,1992). Lastly, PUFAs modulate the expression of nume-rous genes relevant to learning processes (Kitajka et al.,2004) and are integral to cellular inflammatory process-es. Theω-6 fatty acids have proinflammatory properties,as they are precursors to inflammatory eicosanoids (leu-kotrienes, thromboxanes and prostaglandins). Conse-quently, higher ω-6:ω-3 ratios create an environmentconducive to cellular inflammation. The ω-3 PUFAsappear to have significantly less inflammatory effectsand have been viewed by some as anti-inflammatoryagents (Holub, 2002). Overactive or exaggerated im-mune system responses have been implicated in a varietyof systemic and psychiatric disorders, including majordepression (Pearson et al., 2003).

It is worth noting that particular aspects of person-ality, such as negative affect, hostility and cynicism,have been associated with increased disease risk, poorhealth outcomes and increased mortality (Adler andMatthews, 1994; Harris et al., 2003; Russo et al., 1997).With respect to cardiovascular disease, the AmericanHeart Association recommends that all adults increasetheir consumption of fish (a dense source of ω-3 PUFA)to two servings per week (Kris Etherton et al., 2003).This recommendation is based upon strong epidemio-logical and clinical trials evidence that ω-3 fatty aciddeficiency is a reversible risk factor for coronary heartdisease (He et al., 2004; Krauss et al., 2001). While thecardiovascular benefit of increasing ω-3 intakes is wellrecognized, relatively little is known of potential mentalhealth effects among the general public. Likewise, itremains unclear whether the ω-3 and 6 fatty acids dif-ferentially affect mental health. Accumulating evidencegenerally supports a positive and protective role of ω-3fatty acids in psychiatric disorders and suggests morenegative effects of ω-6 fatty acids, particularly AA.Mechanistically, however, little direct evidence supportsthis suggestion. In vitro, the ω-3 fatty acid, DHA, hasbeen shown to uniquely promote growth of hippocam-pal neurons (Calderon and Kim, 2004), while otherevidence suggests that both ω-3 and ω-6 fatty acids canstimulate neurite growth (Darios and Davletov, 2006).Also, decreased levels of ω-6 PUFAs, including AA,have been found in brain and blood of patients withschizophrenia (Skosnik and Yao, 2003).

8 S.M. Conklin et al. / Psychiatry Research 152 (2007) 1–10

In summary, we found evidence that the associationbetween low consumption of ω-3 PUFAs and psycho-logical and behavioral disorders may extend to norma-tive variability in mood, personality and impulsecontrol. Nonetheless, cautious interpretation is neces-sary due to the investigation's small sample, lack ofconsistency amongst some associations, and multiplestatistical comparisons. In as much as, increased fishconsumption is already widely promoted for coronaryheart disease prevention, further research regarding thepsychological and behavioral effects of such dietarychanges is warranted.

Acknowledgements

This research was supported by Public Health Ser-vice Grants HL46328 and HL07560.

References

Adams, P.B., Lawson, S., Sanigorski, A., Sinclair, A.J., 1996. Arachi-donic acid to eicosapentaenoic acid ratio in blood correlatespositivelywith clinical symptoms of depression. Lipids 31, 157–161.

Adler, N., Matthews, K., 1994. Health psychology: why do somepeople get sick and some stay well? Annual Review of Psychology45, 229–259.

Anderson, J.W., Johnstone, B.M., Remley, D.T., 1999. Breast-feedingand cognitive development: a meta-analysis. American Journal ofClinical Nutrition 70, 525–535.

Arab, L., 2003. Biomarkers of fat and fatty acid intake. Journal ofNutrition (Suppl 130), 925S–932S.

Barratt, E.S., 1985. Impulsiveness defined within a systems modelof personality. In: Spielberger, C.D., Butcher, J.M. (Eds.),Advances in Personality Assessment. Lawrence Earlbaum, Hills-dale, NJ, pp. 113–132.

Barratt, E.S., 1994. Impulsiveness and aggression. In: Monahan, J.,Steadman, H.J. (Eds.), Violence and Mental Disorder: Develop-ments in Risk Assessment. University of Chicago Press, Washing-ton DC, pp. 61–79.

Beck, A.T., Steer, R.A., Garbin, M.G., 1988. Psychometric propertiesof the Beck Depression Inventory: twenty-five years of evaluation.Clinical Psychology Review 8, 77–100.

Buydens-Branchey, L., Branchey, M., McMakin, D.L., Hibbeln, J.R.,2003a. Polyunsaturated fatty acid status and aggression in cocaineaddicts. Drug and Alcohol Dependence 71, 319–323.

Buydens-Branchey, L., Branchey, M., McMakin, D.L., Hibbeln, J.R.,2003b. Polyunsaturated fatty acid status and relapse vulnerabilityin cocaine addicts. Psychiatry Research 120, 29–35.

Calder, P.C., Grimble, R.F., 2002. Polyunsaturated fatty acids,inflammation and immunity. European Journal of ClinicalNutrition 56 (Suppl 3), S14–S19.

Calderon, F., Kim, H.Y., 2004. Docosahexaenoic acid promotesneurite growth in hippocampal neurons. Journal of Neurochem-istry 90, 979–988.

Chiu, C.C., Huang, S.Y., Su, K.P., Lu, M.L., Huang, M.C., Chen, C.C.,Shen, W.W., 2003. Polyunsaturated fatty acid deficit in patientswith bipolar mania. European Neuropsychopharmacology 1,99–103.

Clandinin, M.T., Suh, M., Hargreaves, K., 1992. Impact of dietaryfatty acid balance on membrane structure and function of neuraltissues. In: Bazan, N.G., et al. (Eds.), Neurobiology of EssentialFatty Acids. Plenum Press, New York.

Costa, P.T., McCrae, R.R., 1992. Revised NEO Personality Inventory(NEO-PI-R) and NEO Five-factor Inventory (NEO-FFI): Profession-al Manual. Psychological Assessment Resources, Inc., Odessa, FL.

Darios, F., Davletov, B., 2006. Omega-3 and omega-6 fatty acidsstimulate cell membrane expansion by acting on syntaxin 3. Nature440, 813–817.

Das, U.N., 2003. Long-chain polyunsaturated fatty acids in the growthand development of the brain and memory. Nutrition 19, 62–65.

Daviglus, M.L., Stamler, J., Orencia, A.J., Dyer, A.R., Liu, K.,Greenland, P., Walsh, M.K., Morris, D., Shekelle, R.B., 1997. Fishconsumption and the 30-year risk of fatal myocardial infarction.New England Journal of Medicine 336, 1046–1053.

De la Pressa Owens, S., Innis, S.M., 2003. Docosahexaenoic andarachidonic acid prevent a decrease in dopoaminergic andserotonergic neurotransmitters in frontal cortex caused by linoleicand alpha-linoleic acid deficient diet in formula-fed piglets. Journalof Nutrition 129, 2088–2093.

Edwards, R., Peet, M., Shay, J., Horrobin, D., 1998. Omega-3polyunsaturated fatty acid levels on the diet and in red blood cellmembranes of depressed patients. Journal of Affective Disorders48, 149–155.

Eldho, N.V., Feller, S.E., Trisham-Nagle, S., Polozova, I.V., Gawrisch,K., 2003. Polyunsaturated docosahexaenoic vs docosapentaenoicacid-differences in matrix properties from the loss of one doublebond. Journal of the American Chemical Society 125, 6409–6421.

Farquharson, J., Cockburn, F., Patrick, W.A., Jamieson, E.C., Logan,R.W., 1992. Infant cerebral cortex phospholipid fatty-acidcomposition and diet. Lancet 340, 810–813.

Fontani, G., Corradeschi, F., Felici, A.,Alfatti, F.,Migliorini, S., Lodi, L.,2005. Cognitive and physiological aspects of omega-3 polyunsatu-rated fatty acid supplementation in healthy subjects. EuropeanJournal of Clinical Investment 35, 691–699.

Frasure-Smith, N., Lesperance, F., Julien, P., 2004. Major depression isassociated with lower omega-3 fatty acid levels in patients with recentacute coronary syndromes. Biological Psychiatry 55, 891–896.

Freeman, M.P., 2000. Omega-3 fatty acids in psychiatry: a review.Annals of Clinical Psychiatry 12, 159–165.

Gesch, C.B., Hammond, S.M., Hampson, S.E., Eves, A., Crowder, M.J.,2002. Influence of supplementary vitamins, minerals and essentialsfatty acids on the antisocial behavior of young adult prisoners. BritishJournal of Psychiatry 181, 22–28.

Green, P., Hermesh, H., Monselise, A., Marom, S., Presburger, G.,Weizman, A., 2006. Red cell membrane omega-3 fatty acids aredecreased in nondepressed patients with social anxiety disorder.European Neuropsychopharmology 16, 107–113.

Hallahan, B., Garland, M., 2004. Essential fatty acids and their role inthe treatment of impulsivity disorders. Prostaglandins, Leuko-trienes, and Essential Fatty Acids 71, 211–216.

Hamazaki, T., Sawazaki, S., Naago, Y., Naga, Y., Kanagawa, Y.,Yazawa, K., 1999. Administration of docosahexaenoic acidinfluences behavior and plasma catecholamine levels at times ofpsychological stress. Lipids 34, s33–s37.

Hamazaki, T., Thienprasert, A., Kheovichai, K., Samuhaseneetoo, S.,Nagasawa, T., Watanabe, S., 2002. The effect of docosahexaenoicacid on aggression in elderly Thai subjects— a placebo-controlleddouble-blind study. Nutritional Neuroscience 5, 37–41.

Hamazaki, T., Sawazaki, S., Itomura, M., Asaoka, E., Nagao, Y.,Nishimura, N., Yazawa, K., Kuwamori, T., Kobayashi, M., 2005.

9S.M. Conklin et al. / Psychiatry Research 152 (2007) 1–10

The effect of docosahexaenoic acid on aggression in young adultsa placebo controlled double-blind study. Journal of ClinicalInvestment 97, 1129–1134.

Harris, J.I., Hibbeln, J.R., Mackey, R.H., Muldoon, M.F., 2004. Statintreatment alters serum n-3 and n-6 fatty acids in hypercholester-olemic patients. Prostaglandins, Leukotrienes, and Essential FattyAcids 71, 263–269.

Harris, K.F., Matthews, K.A., Sutton-Tyrrell, K., Kuller, L.H., 2003.Associations between psychological traits and endothelial func-tion in postmenopausal women. Psychosomatic Medicine 65,402–409.

He, K., Song, Y., Daviglus, M.L., Liu, K., Horn, L., Dyer, A.R.,Greenland, P., 2004. Accumulated evidence on fish consumptionand coronary heart disease mortality. A meta-analysis of cohortstudies. Circulation 109, 2705–2711.

Hibbeln, J.R., 1998. Fish consumption and major depression. Lancet351, 1213.

Hibbeln, J.R., 2002. Seafood consumption, the DHA content ofmothers milk and prevalence rates of postpartum depression: across-national, ecological analysis. Journal of Affective Disorders69, 15–29.

Hibbeln, J.R., Salem Jr., N., 1995. Dietary polyunsaturated fatty acidsand depression: when cholesterol does not satisfy. AmericanJournal of Clinical Nutrition 62, 1–9.

Hjartaker, A., Lund, E., Bjerve, K.S., 1997. Serum phospholipids fattyacid composition and habitual intake of marine foods registered bya semi-quantitative food frequency questionnaire. EuropeanJournal of Clinical Nutrition 51, 736–742.

Holub, B.J., 2002. Clinical nutrition:4. omega-3 fatty acids incardiovascular care. JAMC 166, 608–615.

Huan, M., Hamazaki, K., Itomura, M., Liu, H., Kang, W., Watanabe,S., Terasawa, K., Hamazaki, T., 2004. Suicide attempt and n-3 fattyacids levels in red blood cells a case control study in China.Biological Psychiatry 56, 490–496.

Hulbert, A.J., Turner, N., Storlien, L.H., Else, P.L., 2005. Dietary fatsand membrane function: implications for metabolism and disease.Biology Review 80, 155–169.

Iribarren, C., Markovitz, J.H., Jacobs, D.R., Schreiner, P.J., Daviglus,M., Hibbeln, J.R., 2004. Dietary intake of n-3, n-6 fatty acids andfish: relationship with hostility in young adults — the CARDIAstudy. European Journal of Clinical Nutrition 58, 24–31.

Itomura, M., Hamazaki, K., Sawazaki, S., Kobayashi, M., Terasawa, K.,Watanabe, S., Hamazaki, T., 2005. The effect of fish oil on physicalaggression in schoolchildren — a randomized, double-blind,placebo-controlled trial. Journal of Nutrition Biochemistry 16,163–171.

Kitajka, K., Sinclair, A.J., Weisinger, R.S., Weisinger, H.S., Mathai, M.,Jayasooriya, A.P., Halver, J.E., Puskas, L.G., 2004. Effects of dietaryomega 3 polyunsaturated fatty acids on brain gene expression.Proceedings of the National Academy of Sciences of the UnitedStates of Americas 101, 10931–10936.

Kris Etherton, P.M., Harris, W.S., Appel, L.J., 2003. Omega 3 fattyacids and cardiovascular disease — new recommendations fromthe American Heart Association. Arteriosclerosis Thrombosis andVascular Biology 23, 151–152.

Krauss, R.M., Eckel, R.H., Howard, B., Appel, L.J., Daniels, S.R.,Deckelbaum, R.J., et al., 2001. Revision 2000: a statement forhealthcare professionals from the nutrition committee of theAmerican Heart Association. Journal of Nutrition 131, 132–146.

Logan, A.C., 2003. Neurobehavioral aspects of omega-3 fatty acids:possible mechanisms and therapeutic value in major depression.Alternative Medicine Review 8, 410–425.

Maes, M., Chrisophe, A., Delanghe, J., Altamura, C., Neels, H.,Meltzer, H.Y., 1999. Lowered ω3 polyunsaturated fatty acids inserum phospholipids and cholesteryl esters of depressed patients.Psychiatry Research 85, 275–291.

Mamalakis, G., Tornaritis, M., Kafatos, A., 2002. Depression andadipose essential polyunsaturated fatty acids. Protaglandins,Leukotrienes and Essentential Fatty Acids 67, 311–318.

Manuck, S.B., Flory, J.D., Ferrell, R.E., Mann, J.J., Muldoon, M.F.,2000. A regulatory polymorphism of the monoamine oxidase—Agene may be associated with variability in aggression, impulsivity,and central nervous system serotonergic responsivity. PsychiatryResearch 95, 9–23.

Marangell, L.B., Martinez, J.M., Zboyan, H.A., Seung Kim, H.F.,Puryear, L.J., 2003. A double-blind, placebo-controlled studyof the omega-3 fatty acid docosahexaenoic acid in the treat-ment of major depression. American Journal of Psychiatry 160,996–998.

Muldoon, M.F., Ryan, C.M., Sereika, S.M., Flory, J.D., Manuck, S.B.,2004. Randomized trial of the effects of simvastatin on cognitivefunctioning in hypercholesterolemic adults. American Journal ofMedicine 117, 823–829.

Nemets, B., Stahl, Z., Belmaker, R.H., 2002. Addition of omega-3fatty acids to maintenance treatment for recurrent unipolardepressive disorder. American Journal of Psychiatry 15, 477–479.

Ness, A.R., Gallacher, J.E., Bennett, P.D., Gunnell, D.J., Rogers, P.J.,Kessler, D., Burr, M.L., 2003. Advice to eat fish and mood: arandomized controlled trial of men with angina. NutritionalNeuroscience 6, 63–65.

Patton, J.H., Stanford, M.S., Barratt, E.S., 1995. Factor structure of theBarratt Impulsiveness Scale. Journal of Clinical Psychology 5,768–774.

Pearson, T.A., Mensah, G.A., Alexander, R.W., Anderson, J.L.,Cannon, R.O., Criqui, M., 2003. Markers of inflammation andcardiovascular disease application to clinical and public healthpractice a statement for healthcare professionals from the centersfor disease control and prevention and the American HeartAssociation. Circulation 107, 499.

Peet, M., Horrobin, D.F., 2002. A dose-ranging study of the effects ofethyl-eicosapentaenoate in patients with ongoing depressiondespite apparently adequate treatment with standard drugs.Archives of General Psychiatry 159, 913–919.

Peet, M., Stokes, C., 2005. Omega-3 fatty acids in the treatment ofpsychiatric disorders. Drug 65, 1051–1059.

Peet, M., Murphy, B., Shay, J., Horrobin, D., 1998. Depletion ofomega-3 fatty acid levels in red blood cell membranes ofdepressive patients. Biological Psychiatry 43, 315–319.

Rapoport, S.I., Chang, M.C., Spector, A.A., 2001. Delivery andturnover of plasma-derived essential PUFAs in mammalian brain.Journal of Lipid Research 42, 678–685.

Richardson, A.J., Puri, B.K., 2002. A randomized double-blind,placebo controlled study of the effects of supplementation withhighly unsaturated fatty acids on ADHD related symptoms inchildren with specific learning difficulties. Progress in Neuro-Psychopharmacology and Biological Psychiatry 26, 233–239.

Richardson, A.J., Montgomery, P., 2005. The Oxford-Durham Study: arandomized, controlled trial of dietary supplementation with fattyacids in children with developmental coordination disorder.Pediatrics 115, 1360–1366.

Russo, J., Katon, W., Lin, E., Von Korff, M., Bush, T., Simon, G.,Walker, E., 1997. Neuroticism and extraversion as predictors ofhealth outcomes in depressed primary care patients. Psychoso-matics 38, 339–348.

10 S.M. Conklin et al. / Psychiatry Research 152 (2007) 1–10

Salem Jr., N., Litman, B., Kim, H.Y., Gawrisch, K., 2001. Mechanismsof action of docosahexaenoic acid in the nervous system lipids 36,945–959.

Skosnik, P.D., Yao, J.K., 2003. From phospholipid and fatty aciddefects to altered neurotransmission: is arachidonic acid a nexus inthe pathophysiology of schizophrenia? Prostaglands Leuko.Essenential Fatty Acids 69, 367–384.

Song, C., Li, X., Leonard, B.E., Horrobin, D.F., 2003. Effects ofdietary n-3 or n-6 fatty acids on interleukin-1B-induced anxiety,stress, and inflammatory responses in rats. Journal of LipidResearch 44, 1984–1991.

Stevens, L.J., Zentall, S.S., Abate, M.L., Kuczek, T., Burgess, J.R.,1996. Omega-3 fatty acids in boys with behavior, learning andhealth problems. Physiology and Behavior 59, 915–920.

Stoll, A.L., Severus, W.E., Freeman, M.P., Rueter, S., Zboyan, H.A.,Diamond, E., et al., 1999. Omega-3 fatty acids in bipolar disorder:a preliminary double-blind, placebo-controlled trial. Archives ofGeneral Psychiatry 56, 407–412.

Su, K.P., Huang, S.Y., Chiu, C.C., 2003. Omega-3 fatty acids in majordepressive disorder; a preliminary double-blind, placebo-con-trolled trial. European Neuropsychopharm 13, 267–271.

Tanskanen, A., Hibblen, J.R., Tuomilehto, J., Uutela, A., Haukala, A.,Viinamaki, H., et al., 2001a. Fish consumption and depressivesymptoms in the general population in Finland. PsychiatricServices 52, 529–531.

Tanskanen, A., Hibbeln, J.R., Hintikka, J., Haatainene, K., Honka-lampi, K., Viinamaki, H., Stoll, A.L., 2001b. Fish consumption,depression, and suicidality in a general population. Archives ofGeneral Psychiatry 58, 512–513.

Yao, J.K., Reddy, R.D., 2003. Membrane pathology in schizophrenia:implication for arachidonic acid signaling. Current MedicinalChemistry — Central Nervous System Agents 3, 57–65.

Zanarini, M.C., Frankenburg, F.R., 2003. Omega-3 fatty acid treatmentof women with borderline personality disorder: a double-blind,placebo-controlled pilot study. American Journal of Psychiatry160, 167–169.