To determine and localize the regions within the visual cortex (e.g. primary versus secondary visual cortex) that were most consistently activated, an activation likelihood analysis (ALE) was performed on the subset of the 28 studies, namely the ones that detected significant occipital lobe activity to drug versus neutral cues (n=24). The coordinates of the peak activity within the occipital cortex for these 24 studies were entered into an ALE analysis. Only the clusters in the visual cortex were entered into the analysis, as this investigation was not designed as a comprehensive analysis of all brain regions implicated in drug cue-reactivity. ALE was performed with the GingerALE software (Turkeltaub et al., 2002), which uses an updated ALE algorithm for minimizing within-experiment and within-group effects (Eickhoff et al., 2012). Likelihood estimation values were computed for all focal locations from contributing contrasts. The null distribution statistic from the likelihood estimation was calculated with full-width at half maximum values empirically determined by the sample size of each contributing study. Values were then subjected to a false discovery rate (FDR) of p < .01 with a minimum extent threshold cluster threshold of 100 mm3.
