Does human functional brain organization shift from diffuse to focal with development?

In this issue, Durston and colleagues (Durston, Davidson, Tottenham, Galvan, Spicer, Fossella & Casey, this issue) set out to test a purported developmental shift from ‘diffuse’ to ‘focal’ brain activity during performance of a response inhibition task. As a second point of interest, they aim to compare age-effects measured between two groups (i.e. cross-sectional) with those detected using two measures of the same group at different ages (i.e. longitudinal). Addressing the potential benefits of longitudinal versus cross-sectional studies in developmental functional neuroimaging is timely and important. In this brief commentary, we discuss these two topics from the perspective afforded by our own work on cognitive development using functional MRI. Many functional neuroimaging studies report relatively larger spatial extents (manifest as a greater number of regions, and/or larger regional volumes) of activation in children relative to adults. These differences are commonly referred to as reflecting a change from ‘diffuse’ or ‘distributed’ to ‘focal’ or ‘enhanced’ activity (e.g. Casey, Cohen, Jezzard, Turner, Noll, Trainor, Giedd, Kaysen, Hertz-Pannier & Rapoport, 1995; Casey, Trainor, Orendi, Schubert, Nystrom, Giedd, Castellanos, Haxby, Noll, Cohen, Forman, Dahl & Rapoport, 1997; Casey, Thomas, Davidson, Kunz & Franzen, 2002; Gaillard, Hertz-Pannier, Mott, Barnett, LeBihan & Theodore, 2000; Hertz-Pannier, Gaillard, Mott, Cuenod, Bookheimer, Weinstein, Conry, Papero, Schiff, Le Bihan & Theodore, 1997; Holland, Plante, Weber Byars, Strawsburg, Schmithorst & Ball, 2001; Passarotti, Paul, Bussiere, Buxton, Wong & Stiles, 2003; Stiles, Moses, Passarotti, Dick & Buxton, 2003). Although these terms are frequently used in the literature, an examination of available evidence suggests a need for greater conceptual clarity. Consideration of several issues may be worthwhile in order to determine whether developmental ‘focalization’ occurs as a principle of developmental brain change. First, consistent and testable definitions of the concepts of ‘diffuse’, ‘distributed’, ‘focal’, ‘enhanced’, etc. are required. These definitions should be specific with regard to phenomena that are readily measured using functional neuroimaging, including the number, location, spatial extent and signal magnitude of activated regions. Second, technical and methodological factors known to contribute to ‘focal’ and ‘diffuse’ image attributes should be considered. For example, children and adults tend to differ by task accuracy and response times. Interestingly, it has been demonstrated that activation maps that include task performance errors appear more ‘diffuse’ than maps created without incorrect responses, showing qualitatively greater spatial extents of activation and lower peak signal magnitudes (Murphy & Garavan, 2004). Many studies that report more diffuse activation of children’s brains use blocked fMRI designs with significant group performance differences. Such studies are unable to differentiate between response errors in making activation maps and group comparisons. Even after removing all error trials, studies using rapid event-related designs still find regional