No evidence for the modulation of the readiness potential by respiratory phase

Correlations in science typically matter as indications of underlying causal structures. In two recent studies, a direct correlation between respiration and the cortical readiness potential (RP) is claimed, with support for causation suggested by the use of terms such as ‘impact’ and ‘modulate’. Following these papers, the notion of a causal relation between respiration and the RP has been repeated in the literature. However, to support the claim that respiratory phase is directly correlated with RP amplitude, all reasonable potential confounds must first be ruled out. We show that a simple confound suffices to explain the observed coupling between respiration phase and the RP: RP amplitude is also coupled with movement onset, which, as these studies rightly reveal, tends to coincide with breathing phase. To test for direct correlation, we must fix one of the variables and test its effect on the other. Grouping RP amplitudes according to respiratory phase (thereby fixing the respiration-behavior coupling), we show that evidence for an RP-respiration coupling disappears. We show this to be true in the original dataset, in a new independent dataset we collected, and in a simulated dataset expressly engineered to have no direct coupling.