Finally, smoking habits significantly correlated with P3 amplitude. However, supplemental HMRs without inclusion of the tobacco variable show that smoking per se was not a statistically significant confound. The supplemental analyses were highly consistent with the main analysis in which smoking was included.
