← Back to Search
Selection Of Cognitive Tests For Trials Of Therapeutic Agents.
Published 2016 · Medicine
Reduce the time it takes to create your bibliography by a factor of 10 by using the world’s favourite reference manager
Time to take this seriously.
We read with interest Carrie Shilyansky and colleagues’ Article concerning the effects of anti-depressant therapy on cognition. That this topic should receive attention in The Lancet Psychiatry is extremely welcome. However, we urge caution with respect to the authors’ contention that the treatments assessed defi nitively had no positive eff ects. In our opinion, a key limitation of this claim was the unreliable nature of the measures used to assess cognition. The recorded test–retest improvements were very substantial, and the reliability of some of the tests were “unacceptably poor”, as Richard Keefe highlights in his Comment. We also wonder whether the authors can legitimately claim to have assessed key domains of function. For example, it is surprising to read of “forward digit span” being used as a measure of working memory. Ordinarily, one would expect the selected test to require the cognitive manipulation of the items to be remembered in order for it to qualify as a measure of working memory, as occurs in backward digit span. Moreover, we were puzzled by the labelling of “decision speed” and “cognitive flexibility” as domains of function. The term domain is usually reserved for cognitive elements that can be clearly dissociated—eg, through the reporting of lesion studies. It seems to us that these constructs are best considered as outcome measures. A further example is “motor coordination”, which was mentioned in the procedures section but not reported thereafter. It is possible that the authors report performance on the tapping task as “psychomotor response speed”; however, if this is the case, a concern is whether a tapping task is a helpful measure. We are sympathetic to the challenges of establishing reliable measures of cognition, but would regard the use of reliable measures as an essential element in the determination of possible treatment effects. The selection of measures that are both stable and reliable is also a key requirement. A helpful precaution is stabilisation of performance through pre-baseline test exposure. Additional methodological consider ations are that the study is not designed primarily to assess cognitive function and that neither the full IntegNeuro battery, nor its component measure, has shown assay sensitivity to potential procognitive effects in placebo-controlled trials of antidepressants in depressive disorders. It would have been helpful to have further information regarding the timings of the various assessments, so that patterns of performance across various measures could be compared. The inclusion of insufficient measures is a substantial limitation to study interpretation and treatment development in psychiatry. Although it might be that the assessed antidepressants have no positive effects on cognition in depression, we are unconvinced that Shilyansky and colleagues’ study represents a categorical demonstration of this fact.