Reply

Article Outline

• References
• Copyright

To the Editor:

Howie's objection to Sallis' characterization of the literature as “a large body of literature” is not merited. In evaluating the regular impact of student participation in exercise on their brain function, the strongest possible causal inferences would benefit from a randomized, controlled trial (RCT), to be sure. The high cost of RCTs makes such evidence relatively rare, unfortunately. In the absence of RCT evidence, however, investigators typically look to available observational data to see whether observed patterns are consistent with there being a beneficial impact of regular student participation in physical activity on their brain function. Moreover, the results of animal studies and studies of the elderly that link exposure to regular exercise to subsequent brain function are also suggestive when they converge with the observational data collected on students. When reviewing these disparate sources of evidence, one finds across >1000 studies convergent evidence that regular physical activity is consistently associated with improved cognition, memory, emotional well-being, attention, time on task, etc., compared with a no-physical activity control. Even when one limits the data to RCTs in this larger literature, Smith et al¹ identified 29 trials, with the preponderance of the trials showing (small) benefits for attention and processing speed, executive function, and memory. Sallis' characterization of the literature as a “large body of literature” is therefore defensible, although he should have made clear what literature he was referring to.

Part of the benefit of RCTs is that the effects of uncontrolled, third-order variables are randomly distributed across conditions. Reliance by Roberts et al² on observational data leave their interpretations open to challenges of the kind enumerated by Howie. They did acknowledge this weakness, inherent in analyses of all observational data when they wrote,

The cross-sectional nature of the data militates against drawing causal inferences from the observed relationship between variations in fitness and variations in standardized test performance. Longitudinal research is needed to confirm whether changes in body composition or physical fitness over time explain variations in test scores and to shed light on the temporal mechanisms that may explain how physical fitness and body composition influence school children's performance on standardized tests.

Howie is right that unobserved influences may explain both the increased fitness and the increased math and reading scores reported by Roberts et al,² but such influences that she cited as parental involvement, overall health status, and school quality are all relevant influences in better understanding the possible connection between exercise and brain function.

Howie asks, “If there has been an overwhelming body of evidence since the mid-1990s, why do policymakers continue to cut physical education and physical activity from schools?” This is certainly not the first time that policymakers have expediently ignored the preponderance of the evidence. It may help to ask a similar question in another domain in which the evidence is more clear cut: “Why do policymakers (in most states, but not in California) continue to allow parents to smoke cigarettes in their cars when their minor age children are present in the car?” More than 3000 peer-reviewed scientific studies in the medical literature have documented the increased risk of bronchitis and asthma exacerbations in children exposed to secondhand smoke, yet policymakers in most states have taken no action to protect the children of parents who smoke by banning smoking in cars when children are passengers. This is just one of thousands of examples, unfortunately, in which policymakers fail to adopt policies reflecting the preponderance of the evidence.

Back to Article Outline

References 

1. Smith PJ, Blumenthal JA, Hoffman BM, et al. Aerobic exercise and neurocognitive performance: a meta-analytic review of randomized controlled trials. Psychosom Med. 2010;72:239–252
   • View In Article
   • CrossRef
2. Roberts CK, Freed B, McCarthy WJ. Low aerobic fitness and obesity are associated with lower standardized test scores in children. J Pediatr. 2010;156:711–718
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (452 KB)
   • CrossRef