High-Order Interdependencies in the Aging Brain

Abstract

Brain interdependencies can be studied either from a structural/anatomical perspective ("structural connectivity", SC) or by considering statistical interdependencies ("functional connectivity", FC). Interestingly, while SC is typically pairwise (white-matter fibers start in a certain region and arrive at another), FC is not; however, most FC analyses focus only on pairwise statistics and neglect highorder interactions. A promising tool to study high-order interdependencies is the recently proposed O-Information, which can quantify the intrinsic statistical synergy and redundancy in groups of three or more interacting variables. In this paper we used the O-Information to investigate how high-order statistical interdependencies are affected by age. For this, we analised functional magnetic resonance imaging (fMRI) data at rest obtained from 164 healthy participants, ranging from 10 to 80 years old. Our results show that older subjects (age ranging from 60 to 80 years) exhibit a higher predominance of redundant dependencies than younger subjects; moreover, this effect seems to be pervasive, taking place at all interaction orders. Additionally, we found that these effects are highly heterogeneous across brain regions, and suggest the existence of a "redundancy core" formed by the prefrontal and motor cortices -thus involving functions such as working memory, executive and motor functions. Our methodology to assess high-order interdependencies in fMRI data has unlimited applications. The code to calculate these metrics is freely available.

High-order functional interactions in ageing explained via alterations in the connectome in a whole-brain model

“…2C. The DMF model reproduced the age differences in redundancy and synergy reported in [24]. Moreover, the differences in redundancy between I4 and the rest of groups (I1,I2,I3) are statistically significant at all orders after the multiple comparison correction of the false discovery rate.…”

“…with the values obtained from the groups of younger participants following previous work [24]. This is illustrated in Fig.…”

“…To study the effect of ageing on brain dynamics, we used functional data from different participants divided into four age groups, similar to previous work [24]: I1…”

“…A stronger connectivity between the frontoparietal and dorsal attention networks has been reported as well [3,19,22,23]. Furthermore, recent investigations on high-order functional interactions have shown the predominance of redundant interdependencies in older adults, consistently across scales [24] and particularly relevant for the DMN [25].…”

Information processing dynamics in neural networks of macaque cerebral cortex reflect cognitive state and behavior

“…One of the most significant limitations is that, while multivariate transfer entropy gives a much more “complete” picture of the information transfer dynamics (and appears to better estimate the generative network [33]), it suffers from a terrible curse of dimensionality and will not scale gracefully as the number of neurons being recorded from gets large and/or the number of bins of history increases. Consequently, for researchers interested in the role of redundancies and synergies in time series dynamics, heuristics such as the O-information may be useful [58, 59, 60]. The partial information decomposition (PID) framework suffers from a similar curse of dimensionality as the the mTE algorithm: the number of unique partial information atoms grows super-exponentially with the number of sources.…”