Here's a great article I stumbled across. According to Richard Gray of Scotsman.com, research teams at Dundee and St Andrews apparently found that reading poetry was more helpful from a cognitive standpoint than reading a novel, even if the novel was by Jane Austin!
Well, I'm not entirely sure whether the conclusion can be projected to great novelists, but what may be more reasonably maintained is that poetry, due possibly to its comparatively (linguistically) indeterminate syntax and semantics, results in more cognitive effort. Reading poetry is to reading novels as hiking up a mountain is to walking down a suburbia street corner.
The research teams held that a relative increase in eye movement corresponded to increased cognitive complexity. Reading poetry created faster eye movements, so it ought to be requiring more sophisticated cognitive mechanisms, as compared to reading a novel.
Other noteworthy indications included the fact that participants would return to one or two lines repeatedly while engaged with a poem. Evidently, the explanation is that the comparative vagueness and/or purposeful equivocation of poetic expressions results in mental perplexity as it tries and discards various ways in which to impart sense to the expression(s).
In any event, I'm so glad to have found a reputable empirical study demonstrating that yes, one of the artistic therapeutic methods I've been using for several years indeed does have cognitively verifiable (even if by implication) benefits.