Subject-specific movement parameters were also included as regressors of no interest. Participant-specific parameter estimates (β values) were calculated at each voxel across the brain. The parameter estimates were then entered into a second level random effects analysis, where one-sample t-tests http://www.selleckchem.com/products/Roscovitine.html were applied to every voxel. Initial statistical thresholding was applied using a threshold of p < .001, uncorrected for multiple comparisons. Activations were considered to be statistically significant only if they survived FWE correction at either the peak or cluster level. For a priori anatomical ROIs, FWE correction was applied using small volume

correction ( Frackowiak et al., 2004) within pre-defined anatomical masks (see Section 2.7). Roxadustat concentration Although not our primary interest, given the results of Park et al. (2007), we also looked for adaptation effects. Here we contrasted trials where the two scenes were perceived to be the same with those that were perceived to be different (including both closer and further away), despite the stimuli being physically identical during any one trial. The trials were divided into these two conditions for modelling both the first and second scene presentations. In all other respects, the experimental design was identical to that described above. We first used a

whole-brain analysis to localise regions which displayed an overall adaptation effect between the first and second scene presentations, regardless of condition. We then conducted more in-depth Molecular motor adaptation analyses using ROIs, as described below. The statistical thresholds were identical to those described above. Given the (limited) previous literature on the functional neuroanatomy of BE, our a priori hypothesis was that the HC would be involved in the BE effect, and that the PHC and RSC might also be implicated. Each of these ROIs was manually defined on the normalised group average T1-weighted structural MR image using the Duvernoy anatomical atlases for guidance ( Duvernoy, 1999, 2005). These anatomical ROIs were used in MarsBar (http://marsbar.sourceforge.net) analyses, where a finite impulse response

(FIR) model ( Dale, 1999; Ollinger et al., 2001) was fitted to the data in order to probe the time-course of responses in ROIs. Four time-windows of 2 sec each were modelled, time-locked to the onset of the first scene presentation. These ROIs were also used for small volume correction within SPM. Based on the whole-brain adaptation analysis described above, we determined that early visual cortex (VC) was a target for further ROI-based analyses. We therefore established a VC ROI using a contrast that was orthogonal to the adaptation analysis (i.e., all scenes presented on the first trial only compared to the implicit baseline). This ROI was used for further adaptation analyses, and for the HC–VC dynamic causal modelling (DCM) analysis described below.