The crucial observation is that persistent eigenmodes of network diffusion appear homologous to characteristic atrophy patterns observed in various dementias. selleck products The first (steady-state) eigenmode, whose eigenvalue is zero, is not shown here, because it is relatively uninteresting, varying simply according to region size, in rough correspondence to atrophy

seen in normal aging. In order to ensure that these results are not due to a specific choice of volumetric algorithm or choice of anatomic atlas, we repeated the same study using volumetric data obtained by the FreeSurfer software (Fischl et al., 2002) and a different 86-region atlas (Figure 5). Measured atrophy patterns generally match the cortical atrophy seen using the automated anatomic labeling (AAL) atlas (Figure 4), but exact match is not to be expected due to both methodological as well as ROI size and shape differences. It is important to note, however, that the visual correspondence between eigenmodes and atrophy remains intact, and the former generally agree with classic

AD/bvFTD pathology, which implies that these results are not methodology-specific. We show later (Figures S5 and S6 available online) that our results are also insensitive to inter-subject variability. The second-most persistent mode (Figures 2 and 4, top rows) closely resembles typical Alzheimer’s atrophy in mesial temporal, posterior cingulate, and limbic structures, as well as lateral temporal and dorsolateral frontal cortex (Apostolova et al., 2007 and Thompson et al., 2003). This eigenmode shows strong involvement of the medial and lateral temporal lobes, which CB-839 molecular weight are involved in memory, and the dorsolateral prefrontal cortex, implicated in working memory (Curtis and D’Esposito, 2003). The main fibers connecting these regions are the superior longitudinal fasciculus (SLF), the splenium of corpus callosum, and the cingulum bundle. While agreement is good with our own volumetric findings (Figures 2 and 4, bottom

rows) and excellent with the published MycoClean Mycoplasma Removal Kit literature (see, for instance, Apostolova et al., 2007, Thompson et al., 2003 and Seeley et al., 2009), there were some areas of disagreement with our volumetric findings in the parietal, frontolateral, and frontoinsular areas. We attribute these differences to small sample size and technical limitations of tractography, co-registration, and volumetrics. The third persistent eigenmode (Figures 3 and 5, top rows) is in good agreement with our bvFTD data (Figures 3 and 5, bottom rows) and published findings (Du et al., 2007, Boxer and Miller, 2005 and Seeley et al., 2009), which indicate prominent atrophy in the orbitofrontal and anterior cingulate regions. This eigenmode is particularly strong in the lateral temporal lobe and the superior frontal, dorsolateral, and orbital cortices— areas that deal with executive function, decision making, expectation, balancing risk versus reward, and inhibition.