Corrections are being reported for the article “Brain Reactivity to Smoking Cues Prior to Smoking Cessation Predicts Ability to Maintain Tobacco Abstinence” by Janes et al., which appeared in Biological Psychiatry, Volume 67, Issue 8 (2010; 67:722–729).
The authors recently discovered a bug in the functional magnetic resonance imaging (fMRI) data processing package, Brain Voyager™ QX. The bug results in the inappropriate display of fMRI activation cluster sizes, which affect the correction for multiple comparisons. Brain Voyager™ acknowledged the problem and issued a bug fix (v2.1.2 Maintenance Release) on their website: (http://www.brainvoyager.com/resources/knownissues/knownissues_if.html).
Therefore, the authors corrected the cluster extent calculations, which resulted in a reduced number of regions that met statistical threshold. The corrected analysis and subsequent article changes noted below do not alter the main findings of the report. Text changes are noted using bold font.
Under the fMRI Analysis subheading of the Methods section, the final sentence should read: “Following 10,000 simulations, the cluster size probability was determined and the cluster extent that yielded p = .05 (20 3×3×3mm re-sampled isotropic voxels, or ~540 mm3) was selected…”.
Under the Independent Component Analysis (ICA) subheading of the Methods section the final sentence should read: “Multiple comparisons were corrected to p ≤ .05 with the Monte Carlo procedure described earlier.”
Under the Results section, Functional MRI Results Whole-Brain Analysis subheading, the text should read: “A whole-brain mixed-effects analysis revealed that, relative to the abstinent group, the slip group had greater smoking-related versus neutral image reactivity in the bilateral insula, prefrontal cortex, posterior cingulate, parahippocampal gyrus, putamen, thalamus, cerebellar hemispheres and vermis (cluster corrected to p ≤ .05). Additional areas of activation were noted at a less stringent level of statistical significance and included the ACC, amygdala, primary motor cortex, premotor cortex, inferior parietal cortex, and striate and extrastriate cortex (t19 = 2.86, uncorrected p ≤ .01, cluster extent 31 mm3, Figure 1, Table S1 in Supplement 1).”
Under the Results section, Functional Connectivity Results subheading, the second half of the paragraph should read: “When comparing this network between subjects who slipped and maintained abstinence, slip subjects showed decreased functional connectivity between this network and the left insula, prefrontal cortex, ACC, cerebellum, and middle temporal gyrus (cluster corrected to p ≤ .05). Additional areas of decreased functional connectivity were noted at a less stringent level of statistical significance and included the inferior frontal gyri, primary and premotor cortex, primary somatosensory cortex, superior temporal gyrus, putamen and primary and association visual cortices (t20 = 2.58, uncorrected p ≤ .01, cluster extent 31 mm3, Figure 2, Table S3 in Supplement 1).”
The second sentence of the Figure 1 legend should read: “Subjects who slipped exhibited greater functional magnetic resonance imaging reactivity in several brain areas (Table S1 in Supplement 1), including the right anterior insula [t19 = 2.86, uncorrected p ≤ .01, cluster corrected to p ≤ .05, uncorrected data displayed].”
The Figure 2 legend, bottom panel, last sentence, should read: “Talairach (55) coordinates for the dorsal anterior cingulate cortex: x = 0, y = 13, z = 26, (t20 = 2.58, uncorrected p ≤ .01, cluster corrected to p ≤ .05, uncorrected data displayed in bottom panel).”
For Tables S1 and S3 in Supplement 1, the final sentence of the legend should read: “Voxels refer to the cubic volume of each cluster. Clusters of 540 mm3 or greater are cluster corrected to p ≤ .05.”
For Table S2 in Supplement 1, the final sentence of the legend should read: “Voxels refer to the cubic volume of each cluster.”