The Corpus Callosum

Objective: Key biological factors that influence the development of depression are modified by diet. This study examined the extent to which the high-prevalence mental disorders are related to habitual diet in 1,046 women ages 20-93 years randomly selected from the population.

Method: A diet quality score was derived from answers to a food frequency questionnaire, and a factor analysis identified habitual dietary patterns. The 12-item General Health Questionnaire (GHQ-12) was used to measure psychological symptoms, and a structured clinical interview was used to assess current depressive and anxiety disorders.

Results: After adjustments for age, socioeconomic status, education, and health behaviors, a "traditional" dietary pattern characterized by vegetables, fruit, meat, fish, and whole grains was associated with lower odds for major depression or dysthymia and for anxiety disorders. A "western" diet of processed or fried foods, refined grains, sugary products, and beer was associated with a higher GHQ-12 score. There was also an inverse association between diet quality score and GHQ-12 score that was not confounded by age, socioeconomic status, education, or other health behaviors.

Conclusions: These results demonstrate an association between habitual diet quality and the high-prevalence mental disorders, although reverse causality and confounding cannot be ruled out as explanations. Further prospective studies are warranted.

Unfortunately, the full text is not openly accessible.  The authors found that, after adjusting for potential confounding factors, that "a 1.00-SD increase in the score for traditional dietary pattern was associated with a 35% reduced odds for major depression or dysthymia and a 32% reduced odds for anxiety disorders."  This is a reasonably strong effect size. 

Naturally, studies such as these have plenty of problems that make it difficult to draw firm conclusions.  One problem is that it is tempting to believe the study, merely because if appears to confirm what one might already suspect to be true. The authors point out that reverse causality is a possibility.  That is, it is possible that persons who already are anxious or depressed might tend to have worse dietary practices.  This could lead to the same kind of correlation, but the correlation would be reversed: the poor diet would be caused by the emotional problems, as opposed to the poor diet causing the emotional problems.  Also, even though they tried to correct for confounding variables, it is possible that the statistics corrections did not filter out the confounding effects.

Nutrition and Psychiatry
Freeman
Am J Psychiatry 167 (3): 244

There is no abstract for editorials, so I will provide a brief excerpt.  The purpose of the editorial is to put the Jacka study in context, by citing some additional studies:

Two prospective cohort studies from Spain (6) and the United Kingdom (7) also provide information regarding the relationship between depression and patterns of nutritional intake. In the Spanish study (6), the Mediterranean dietary pattern was found to confer protection against the development of depression. Adherence to the Mediterranean diet was categorized into quartiles, with incremental increases in risk for depression associated with lower adherence. Particular inverse associations with incident depression were noted with higher consumption of fruit, nuts, and legumes, as well as a higher ratio of monounsaturated to saturated fatty acids. In the U.K. study, a dietary pattern containing higher amounts of processed foods was associated with a higher risk of subsequent development of depressive symptoms (7). Dietary patterns were assessed for "whole foods" (with fruit, vegetables, and fish characteristic of intake) and processed foods (largely represented by processed meat and bread products and high-fat dairy products). After adjustment for confounding variables, persons in the group with the highest intake of whole foods had the lowest rate of depressive symptoms on the Center for Epidemiologic Studies Depression scale. In addition, a study from France demonstrated an association between the metabolic syndrome and depression (8).

These three studies tend to confirm the conclusions from the Jacka study.  More fruit, nuts, legumes, fish, and monounsaturated fats = good.  More processed foods and high-fat dairy products = bad.  Metabolic syndrome = bad (although we already know this). 

Although these studies are still not conclusive to the extent that I would like, I feel comfortable saying that persons with anxiety and depression are well-advised to try to improve their food selection practices.