external stimulation. The focus was on the type of interactions
occurring between different brain areas in terms of positive/nega-
tive correlations in one case, and in terms of synchronous activation
in the other.3.1 The Role
of Negative
Correlations
3.1.1 The Problem
Up to now, studies on functional brain networks mainly focused on
positive correlations between cerebral areas due to the still not well-
defined nature of the negative ones. Negative correlations have
been often interpreted as a preprocessing artifact, namely as a global
signal regression [21], although significant relations have been
found between negative correlations and caffeine intake, medita-
tion, brain development and aging, schizophrenia, different social
and cognitive tasks [22]3.1.2 The Model A possible topology of negative functional networks has been
described by several recent papers by means of Network Theory
[23] and of Balance Theory [24].While the former theory is
particularly useful in providing quantitative parameters of network
topologies, the latter one helps in defining the conditions for the
networks’ functional stability. This is of special interest, since it may
suggest possible mechanisms accounting for the influence of nega-
tive correlations on the functional equilibria between brain areas.
A well-known feature of the MAS is the flexible rearrangement
of connection patterns according to simple rules. Such rules, in the
case of networks, smooth the difference in the activity levels of
Fig. 6Correlation matrix between Regions of Interest (ROIs) of a single subject from functional Magnetic
Resonance Imaging (fMRI) data. The raw activation (BOLD effect) data are plotted in a normalized, false color
scale. The original fMRI images of the subject are divided into 90 ROIs and from each ROI the time series are
extracted (acquisition time 450 s at 3 Hz frequency, in order to obtain discrete series of 150 elements). The
correlation matrix was calculated for all possible couples of the 90 ROIs (modified from [32])
316 Alfredo Colosimo