But 1) does this pattern hold at the individual level? And 2) is it just the worst offender -- sugar -- or are carbohydrates in general depressing? Also, 3) what mechanisms may be involved?
1) Yes, among individuals, high soft drink consumption is associated with more behavioral and mental distress problems (free full text). They couldn't tell whether it the effect is due to sugar or something else in soda, like caffeine, but below we'll see that sugar is surely the culprit. Interesting fact from the article: "Norway has the highest soft drink consumption rate in the world, 115 L (245 pt) per inhabitant per year." No wonder those Scandinavians are so gloomy.
2) No, it's not just sugar but carbs in general (free PDF here). If you try to find out how nutrition affects anything, you'll be very frustrated to learn that no one has studied the question, whatever it is -- or if so, that it was so long ago that it doesn't appear in PubMed or within the first 100 google results. You can easily read about how the sexy micronutrient du jour affects everything, but not how simple macronutrients like fats vs. carbs vs. proteins affect anything. Well, you can read about omega-3 vs. omega-6 polyunsaturated fats, but that's it.
Here is the article's review of the "few studies of associations between macronutrients and mood":
Subjects with MDD tend to consume more carbohydrates in their diets than non-depressed individuals (Christensen and Somers, 1996), and they show heightened preference for sweet carbohydrate or fat- rich foods during depressive episodes (Christensen, 2001). High carbohydrate intakes increase brain uptake of the amino acids tryptophan, which in turn stimulates the synthesis of serotonin (Rogers, 2001.) This seems, for example, to rapidly lead to drowsiness in healthy subjects but to alertness in patients with SAD (Rosenthal et al. 1989). A detailed study of the lifestyle of 89 bipolar patients and 445 age- and sexmatched controls showed that total daily sucrose intake, %age of energy from carbohydrate, and consumption of sweetened drinks were higher in bipolar patients (Elmslie et al. 2001). Westover and Marangell (2002) reported a highly significant 23 correlation between sugar consumption and annual rates of depression in a cross-national study involving six countries.
Fats and proteins
In general, a low-fat diet may have negative effects on mood (Wells et al. 1998), and altered dietary fat intake can lead to acute behavioural effects such as drowsiness, independent of energy consumption, in healthy subjects (Lloyd et al. 1994). The intake of branched-chain amino acids may acutely alleviate manic symptoms in patients with adequate drug treatment (Scarnà et al. 2003), and a high intake of proteins also seems to increase alertness (Rogers, 2001).
3) It would seem that glucose intake should make you happier because insulin acts to store sugars and non-tryptophan amino acids. It is tryptophan that's converted into serotonin in the brain, so more glucose creates more insulin creates more available tryptophan (it now has less competition from other amino acids in getting transported into the brain). More tryptophan means more serotonin, which means stabler mood. The fact that chronically high levels of sugar are associated with depression suggests that the short-term vs. long-term effects of glucose on serotonin concentration may be different.
Indeed, that's just what a group of Portuguese researchers found when they studied how much serotonin was taken up (i.e. made unavailable) by Caco-2 cells that were treated with glucose compared to no glucose. These cells express the serotonin transporter; greater activity of the transporter makes less serotonin available where it improves mood.
After short-term exposure to high levels of glucose (2 hours at 40 mM), the cells showed about 25% less uptake of serotonin, meaning more of it was available. So, in the short-term, a temporary bit of glucose can improve your mood.
However, when exposed to high levels of glucose for longer periods -- 21 to 24 weeks -- the reverse happened. Now the glucose-treated cells showed about 30% higher uptake of serotonin, making less of it available. So, if you have chronic high levels of glucose -- i.e., if you eat a carbohydrate-rich diet -- you will show more signs of depression and mental instability.
Finally, there's an animal model too, where rats fed a sugar-rich diet show lower serotonin levels and increased eating (hyperphagia). The researchers here discovered that the effect probably does not involve levels of tryptophan, but rather the activity level of the enzyme that converts tryptophan into serotonin.
Summing it all up, it looks like a shot in the arm of glucose can make you feel better in the short-term, assuming that your diet overall is low in carbs and high in fat. I do this by just eating a single big bite of a date and walnut wedge, or four spoonfuls of coconut milk-based ice cream. (But make sure to have this at a time in the day when you don't need energy -- otherwise the glucose spike will trigger insulin production, which keeps your fatty acid fuels locked in fat cells where you can't use them.) In the Norwegian soda study, they mention that adolescents who didn't consume any sugar were a bit worse than those who had a little bit, and that supports that idea of having just a little bit for short-term mood improvement.
If you follow a low-fat, and thus high-carb diet, however, less serotonin will be available, and you'll end up more moody and unstable. This will be especially pronounced if the carbs are easily digestible like starches and sugars, but eating lots of whole grain rice and wheat bread will get you too. Fortunately, those foods don't taste like anything by themselves, so you're not really giving up anything by chucking them in the trash can. A whole-wheat baguette with lean tuna, olive oil, and pepper doesn't stand a chance against a toasted blue-corn tortilla with a stack of pastrami or pepperoni, a heap of cole slaw, and melted cheese.