Table 1.

Names, p-values, and syntax of conjunctions used for statistical analysis of functional fMRI data.

Condition abbreviations
AlloVisL	Allocentric	–visual feedback	– left visual field cursor movement
AlloNoVisL	Allocentric	–no-visual feedback	– left visual field cursor movement
TarVisL	Target-directed	–visual feedback	– left visual field cursor movement
TarNoVisL	Target-directed	–no-visual feedback	– left visual field cursor movement
AlloVisR	Allocentric	–visual feedback	– right visual field cursor movement
AlloNoVisR	Allocentric	–no-visual feedback	– right visual field cursor movement
TarVisR	Target-directed	–visual feedback	– right visual field cursor movement
TarNoVisR	Target-directed	–no-visual feedback	– right visual field cursor movement
“Experimental task > baseline,” p < 0.005, cluster size threshold = 252 mm3 (AlloVisL > 0) AND (AlloNoVisL > 0) AND (TarVisL > 0) AND (TarNoVisL > 0) AND (AlloVisR > 0) AND (AlloNoVisR > 0) AND (TarVisR > 0) AND (TarNoVisR > 0) “Visual feedback > no-visual feedback,” p < 0.005, cluster size threshold = 72 mm3 (AlloVisL > 0) AND (TarVisL > 0) AND (AlloVisR > 0) AND (TarVisR > 0) AND (AlloVisL > AlloNoVisL) AND (AlloVisR > AlloNoVisR) AND (TarVisL > TarNoVisL) AND (TarVisR > TarNoVisR) “No-visual feedback > visual feedback,” p < 0.005, cluster size threshold = n.a. (AlloNoVisL > 0) AND (TarNoVisL > 0) AND (AlloNoVisR > 0) AND (TarNoVisR > 0) AND (AlloNoVisL > AlloVisL) AND (AlloNoVisR > AlloVisR) AND (TarNoVisL > TarVisL) AND (TarNoVisR > TarVisR) “Allocentric > target-directed,” p < 0.005, cluster size threshold = 99 mm3 (AlloVisL > 0) AND (AlloNoVisL > 0) AND (AlloVisR > 0) AND (AlloNoVisR > 0) AND (AlloVisL > TarVisL) AND (AlloNoVisL > TarNoVisL) AND (Allo- VisR > TarVisR) AND (AlloNoVisR > TarNoVisR) “Allocentric > target-directed (cursor movement in LVF, relevant target-dots in RVF),” p < 0.005, cluster size threshold = 153 mm3 (AlloVisL > 0) AND (AlloNoVisL > 0) AND (AlloVisL > TarVisL) AND (AlloNoVisL > TarNoVisL) “Allocentric > target-directed (cursor movement in RVF, relevant target-dots in LVF),” p < 0.005, cluster size threshold = 117 mm3 (AlloVisR > 0) AND (AlloNoVisR > 0) AND (AlloVisR > TarVisR) AND (AlloNoVisR > TarNoVisR) “Target-directed > allocentric,” p < 0.005, cluster size threshold = n.a. (TarVisL > 0) AND (TarNoVisL > 0) AND (TarVisR > 0) AND (AllNoVisR > 0) AND (TarVisL > AlloVisL) AND (TarNoVisL > AlloNoVisL) AND (TarVisR > AlloVisR) AND (TarNoVisR > AlloNoVisR) “Target-directed > allocentric (cursor movement in LVF, relevant target-dots in LVF),” p < 0.005, cluster size threshold = n.a. (TarVisL > 0) AND (TarNoVisL > 0) AND (TarVisL > AlloVisL) AND (TarNoVisL > AlloNoVisL) “Target-directed > allocentric (cursor movement in RVF, relevant target-dots in RVF),” p < 0.005, cluster size threshold = 36 mm3 (TarVisR > 0) AND (AllNoVisR > 0) AND (TarVisR > AlloVisR) AND (TarNoVisR > AlloNoVisR) “Visual feedback target-directed > visual feedback allocentric,” p < 0.0025, cluster size threshold = 36 mm3 (TarVisR + TarVisL > 0) AND (TarVisR + TarVisL > AlloVisR + AlloVisL) “Visual feedback allocentric > visual feedback target-directed,” p < 0.0025, cluster size threshold = 171 mm3  (AlloVisR + AlloVisL > 0) AND (AlloVisR + AlloVisL > TarVisR + TarVisL)

Individual statistical tests within a conjunction are written in parenthesis. “AND” indicates the logical AND operator. Statistical thresholds (p-values) were those applied to each of the component contrasts in the conjunction (Nichols et al., 2005). Probability of type-I errors in the resulting activation map was controlled using cluster size thresholds (Forman et al., 1995). Cluster size thresholds are computed taking into account the spatial smoothness (or correlation) of an activation map (Forman et al., 1995; Goebel et al., 2006). Therefore, cluster size thresholds have to be determined for each activation map separately. It is for this reason that cluster size thresholds could differ among conjunctions. Cluster size thresholds were not computed when no significant activation was observed. In those cases, cluster size threshold were not applicable (n.a.; compare Table 3). To enable comparison of results across analyses, conjunctions that use simple comparisons (i.e., tests that compare one condition to another) use p < 0.005 for each individual test, whereas conjunctions that use complex comparisons (i.e., tests that compare averages of two conditions) use p < 0.0025.