Defining a functional network homeostasis after stroke: EEG-based approach is complementary to functional MRI

We have read with great interest the article by Adhikari et al . entitled ‘Decreased integration and information capacity in stroke measured by whole brain models of resting state activity’ recently published in Brain (Adhikari et al. , 2017). The authors of this excellent article found that the brain network functional connectivity, evaluated by functional MRI, is impaired among subjects with subacute stroke, being the graph theoretical measures of integration and segregation decreased. This kind of network reorganization occurs both in the whole cerebral network and in the seven resting state subnetworks [dorsal attention network (DAN), ventral attention network (VAN), motor network, visual network, frontal parietal network (FPN), language network (LAN) and default mode network (DMN)]. In particular, the integration is decreased among all resting state net- works, while the segregation, intended as mean information capacity, is decreased globally and among DAN and FPN resting state networks. In this frame, we would like to convey an analogous and complementary EEG-based approach that could add more ‘dynamic’ information about functional cortical connectivity, being EEG signals directly related to the cyclic firing of the neuronal assemblies, reaching thus a temporal discrimin- ation—particularly when fast EEG rhythms are con- sidered—of few tens of milliseconds. Requiring a balance in the brain activity between local specialization and global integration (Tononi et al ., 1994), properly quantified by a small-world network model (Watts and Strogatz, 1998), characterized by high clustering coefficient (index of functional segregation) and short path length coefficient (index of functional integration) (Bassett and Bullmore, 2006; Stam and Reijneveld, 2007), we evaluated via EEG the small-world characteristics (small-worldness) of resting state cortical networks in 30 consecutive patients with acute ischaemic stroke (Caliandro et al. , 2017). In fact, using the eLORETA software (Pascual-Marqui, 2002), it is possible to reconstruct 42 regions of interest, corresponding to 42 Brodmann areas, for each hemisphere, and to calculate the current density time series of the regions of interest (lagged linear coherence) (Pascual-Marqui, 2007; Pascual- Marqui et al ., 2011) between all possible pairs of the regions of interest for each of the seven independent EEG frequency bands of delta (2–4Hz), theta (4–8Hz), alpha 1 (8–10.5Hz), alpha 2 (10.5–13Hz), beta 1 (13–20Hz), beta 2 (20–30Hz), and gamma (30–45Hz) rhythms for each subject. Given this, an EEG-derived cortical network in which the nodes are represented by the Brodmann areas and the edges are weighted by lagged linear connectivity values can be reconstructed (Pascual-Marqui, 2007). The aforementioned EEG- and graph theory-based approach allowed us to find network rearrangement in a frequency-dependent modality doi:10.1093/brain/awx271 BRAIN 2017: 140; 1–2 | e71 Advance Access publication November 3, 2017 ß The Author (2017). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For Permissions, please email: [email protected] Downloaded from https://academic.oup.com/brain/article-abstract/140/12/e71/4590109 by universit� cattolica del sacro cuore user on 20 February 2018 (Caliandro et al. , 2017). In particular, in the delta band we found a decreased small-worldness (similar to the rearrangement in delta band, we found a bilateral theta band small-worldness reduction only in patients with left hemispheric stroke). On the hand, we found an increased small-worldness in alpha 2 band. It is noteworthy that the abovementioned network changes were found in the hemi- sphere ipsilateral to the ischaemic lesion, in the contralateral hemisphere and in the whole brain. Compared with the Adhakiri et