Coordinate‐based activation likelihood estimation meta‐analysis of neuroimaging data: A random‐effects approach based on empirical estimates of spatial uncertainty

Abstract

A widely used technique for coordinate-based meta-analyses of neuroimaging data is activation likelihood estimation (ALE). ALE assesses the overlap between foci based on modelling them as probability distributions centred at the respective coordinates. Here we present a revised ALE algorithm addressing drawbacks associated with former implementations:The first change pertains to the size of the probability distributions, which had to be specified by the used. To provide a more principled solution, we analysed fMRI data of 21 subjects, each normalised into MNI space using nine different approaches. This analysis provided quantitative estimates of between-subject and between-template variability for 16 functionally defined regions, which were then used to explicitly model the spatial uncertainty associated with each reported coordinate. Secondly, instead of testing for an above-chance clustering between foci, the revised algorithm assesses above-chance clustering between experiments. The spatial relationship between foci in a given experiment is now assumed to be fixed and ALE results are assessed against a nulldistribution of random spatial association between experiments. Critically, this modification entails a change from fixed-to random-effects inference in ALE analysis allowing generalisation of the results to the entire population of studies analysed.By comparative analysis of real and simulated data, we showed that the revised ALE-algorithm overcomes conceptual problems of former meta-analyses and increases the specificity of the ensuing results without loosing the sensitivity of the original approach. It may thus provide a methodologically improved tool for coordinate-based meta-analyses on functional imaging data.

Functional Heterogeneity in Posterior Parietal Cortex Across Attention and Episodic Memory Retrieval

“…One possibility is that, during recognition memory decisions, lateral IPS serves as a "mnemonic accumulator" of signals (Eickhoff et al 2009) of relevant studies in the memory and attention literatures (thresholded with a FDR of all P < 0.05, orange; overlap rendered in yellow). Comparison of the findings in the current study to those reported in the literature reveals overlap between (b) the IPS region displaying memory strength effects here and the IPS region consistently associated with familiarity across studies of episodic retrieval (as compiled in Vilberg and Rugg 2008b), (c) the SPL region positively tracking retrieval reaction time here and the SPL region that is consistently associated with top-down attention across studies of attention (as compiled in Hutchinson et al 2009), (d) the AnG region displaying successful recollection effects here and the AnG region consistently associated with recollection across studies of episodic retrieval (as compiled in Vilberg and Rugg 2008b), and (e) the left TPJ region displaying greater activity for Item Only than for Task and Specific Source trials here and the TPJ region consistently associated with the stimulus-driven reflexive orienting of attention across studies of attention (as compiled in Hutchinson et al 2009).…”

“…Comparison of the findings in the current study to those reported in the literature reveals overlap between (b) the IPS region displaying memory strength effects here and the IPS region consistently associated with familiarity across studies of episodic retrieval (as compiled in Vilberg and Rugg 2008b), (c) the SPL region positively tracking retrieval reaction time here and the SPL region that is consistently associated with top-down attention across studies of attention (as compiled in Hutchinson et al 2009), (d) the AnG region displaying successful recollection effects here and the AnG region consistently associated with recollection across studies of episodic retrieval (as compiled in Vilberg and Rugg 2008b), and (e) the left TPJ region displaying greater activity for Item Only than for Task and Specific Source trials here and the TPJ region consistently associated with the stimulus-driven reflexive orienting of attention across studies of attention (as compiled in Hutchinson et al 2009). Note that only left parietal peak-voxel coordinates were used in the generation of the ALE maps, whereas the bootstrapped null distribution to which they were compared necessarily included voxels throughout the brain (Eickhoff et al 2009). This slightly lowered the resulting P values (and consequently produced more liberal estimates of activation in the ALE maps).…”

A meta‐analysis of functional magnetic resonance imaging studies of eye movements and visual word reading

“…The meta‐analysis was performed using the ALE algorithm (Eickhoff et al., 2009, 2012; Turkeltaub et al., 2012) found in the GingerALE2.3 software (http://brainmap.org/ale/; RRID:SCR_014921). In the ALE approach, spatial probability distributions for the foci were modeled at the center of three‐dimensional Gaussian functions and the Gaussian distributions were aggregated across the entire set of experiments to generate a map of consistencies among studies that estimated the likelihood of activation for each voxel—the ALE statistic (Eickhoff et al., 2009).…”

“…In the ALE approach, spatial probability distributions for the foci were modeled at the center of three‐dimensional Gaussian functions and the Gaussian distributions were aggregated across the entire set of experiments to generate a map of consistencies among studies that estimated the likelihood of activation for each voxel—the ALE statistic (Eickhoff et al., 2009). Coordinates reported in the Talairach space were first transformed into the Montreal Neurological Institute brain template using the appropriate transformation algorithms implemented in GingerALE.…”

“…Previous neuroimaging studies have shown that frontoparietal attentional regions play a critical role in eye movements (Corbetta & Shulman, 2002; Simon, Mangin, Cohen, Le Bihan, & Dehaene, 2002), consistent with their function for spatial representation and spatial updating (Merriam, Genovese, & Colby, 2003; Pertzov, Avidan, & Zohary, 2011; Silver & Kastner, 2009). In a recent meta‐analysis, Jamadar, Fielding, and Egan (2013) have used the activation likelihood estimation method (ALE; Turkeltaub et al., 2012; Eickhoff et al., 2009; Eickhoff, Bzdok, Laird, Kurth, & Fox, 2012) to compare the neural networks of prosaccades and antisaccades. At the cortical level, they found that the network of prosaccades includes the primary visual cortex, extrastriate cortex, parietal eye field (PEF, in the posterior parietal cortex), frontal eye field (FEF, in the superior part of the prefrontal gyrus), and supplementary eye field (SEF, in the medial frontal gyrus [MedFG]).…”