Functional Connectivity Change in Response to Stroke Is Comparable Across Species: From Mouse to Man.

Abstract

HomeStrokeVol. 52, No. 9Functional Connectivity Change in Response to Stroke Is Comparable Across Species: From Mouse to Man Free AccessEditorialPDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyRedditDiggEmail Jump toFree AccessEditorialPDF/EPUBFunctional Connectivity Change in Response to Stroke Is Comparable Across Species: From Mouse to Man Gerard R. Hall, MSc, Marcus Kaiser, PhD, FRSB and Tracy D. Farr, PhD Gerard R. HallGerard R. Hall https://orcid.org/0000-0002-5212-7850 School of Life Sciences, University of Nottingham, United Kingdom (G.R.H., T.D.F.). Search for more papers by this author , Marcus KaiserMarcus Kaiser https://orcid.org/0000-0002-4654-3110 NIHR Nottingham Biomedical Research Centre (M.K.), University of Nottingham, United Kingdom. Sir Peter Mansfield Imaging Centre, School of Medicine (M.K.), University of Nottingham, United Kingdom. Shanghai Jiao Tong University, School of Medicine, Shanghai, China (M.K.). Search for more papers by this author and Tracy D. FarrTracy D. Farr Correspondence to: Tracy D. Farr, PhD, School of Life Sciences, University of Nottingham, Medical School, Queen’s Medical Centre, Nottingham, NG7 2UH, United Kingdom. Email E-mail Address: [email protected] https://orcid.org/0000-0002-6781-5226 School of Life Sciences, University of Nottingham, United Kingdom (G.R.H., T.D.F.). Search for more papers by this author Originally published20 Jul 2021https://doi.org/10.1161/STROKEAHA.121.034097Stroke. 2021;52:2961–2963This article is a commentary on the followingTranslating Functional Connectivity After Stroke: Functional Magnetic Resonance Imaging Detects Comparable Network Changes in Mice and HumansOther version(s) of this articleYou are viewing the most recent version of this article. Previous versions: July 20, 2021: Ahead of Print See related article, p 2948Assessing functional connectivity change with neuroimaging is a key method to understand and predict recovery and plasticity after stroke. Animal models offer the potential to obtain mechanistic insight into the processes that underpin these changes, but functional magnetic resonance imaging connectivity approaches have been lagging in preclinical research. If the goal of bench to bedside translation is to be achieved, it is equally important to identify species commonalities as well as differences. Blaschke et al1 addresses this issue by comparing the same functional connectivity outcomes between a cohort of stroke patients and a group of mice with similar strokes. Thirteen patients with moderate unilateral upper extremity deficits underwent resting-state functional magnetic resonance imaging within 1 week of stroke onset. To complement this, experiments were performed in 2 strains of mice 2 weeks following cortical stroke with either photothrombosis or the distal middle cerebral artery occlusion model. Interestingly, despite minor variation in acquisition parameters and balanced differences in data processing, both humans and mice demonstrated an overall increase in functional connectivity when compared with controls.This is fascinating, particularly since most resting-state functional magnetic resonance imaging studies in stroke have demonstrated decreases in functional connectivity. Indeed, decreases in interhemispheric connectivity is a widely reported phenomenon across a variety of different stroke subtypes, reviewed in Baldassarre et al2 and Desowska et al3 and recovery with time is correlated with improvements in sensorimotor function4 and integrity of transcallosal fibres5 and the corticospinal tract.6 This is not isolated to humans, despite the requirement of sedation/anesthesia for preclinical studies. Similar reports have been published in rats as early as 2 hours7 and out to 10 weeks following stroke.8 It is possible the divergent findings may be due to the analysis approach. For example, most resting-state functional magnetic resonance imaging studies have seeded regions of interest, generally sensorimotor areas, whereas functional connectivity correlations were compiled between all brain regions from atlases with similar parcellation across species.1 Moreover, increases in functional connectivity after stroke have been reported with independent component analysis. This is a data-driven approach that does not require a priori hypotheses. It identifies statistically independent patterns of variation, including resting-state networks. A study of 6 patients with unilateral hand weakness 2 weeks poststroke identified 6 networks, and there were double the number of connections among them in stroke patients compared with controls.9 A similar study in internal capsule stroke at 6 months identified several known resting-state networks.10 Increases in functional connectivity were observed in the sensorimotor, visual, auditory, and dorsal attention networks, combined with decreases in frontoparietal networks of stroke patients compared to controls. Collectively, this suggests additional advantages to a global approach to identify widespread changes in connectivity patterns.Blaschke et al1 employed graph theory to interrogate connectivity network characteristics. This mathematical approach designates anatomic regions as nodes and connectivity is depicted by relationships between them (edges). Neural networks have consistent features such as modules, hubs, and a hierarchical organization that is evident across a wide range of species.11 Interestingly, the observed overall increase in functional connectivity was accompanied by increases in mean nodal strength and small-worldness, a well-known property of brain networks. On the surface, this may imply that stroke brains outperform controls, but many different network topologies contribute to small-worldness, and the term reflects a balance between high nodal clustering (a measure of segregation and specialization) and short average internodal path lengths (a measure of integration and efficiency). The increase in small-worldness was driven by reductions in path lengths in both species, and increases in the clustering coefficient in mice (Figure). These findings compare to another study that acutely imaged 29 proximal anterior circulation stroke patients, and observed reduced path lengths without significant differences in clustering compared to age-matched controls.12 This pattern was evident across a range of thresholds and suggests a surge in interactions between brain regions, in particular, an increase in the number of long-range connections.13 However, there are additional studies that suggest this explanation is not so straightforward. Rats with stroke exhibited increases in clustering and path lengths compared with controls acutely, suggesting a less integrated and more segregated network.14 Ultimately, this study provided a dynamic picture of the recovering brain network with time; imaging was repeated out to 70 days. Although both parameters normalized over time in rats with small subcortical strokes, animals with large lesions maintained increases in path lengths while clustering decreased with time. Decreases in clustering insinuate that the network became more random.Download figureDownload PowerPointFigure. Depiction of a hypothetical functional connectome in which stroke affects 3 nodes and 8 edges, and this results in increased clustering and decreased path lengths.Comparably, a group of 10 subcortical stroke patients with motor deficits was imaged over the course of a year, and clustering decreased without any change in path lengths.15 Interestingly, the first scan was acquired 1 week after stroke, and there were no differences in either parameter compared to controls.Dynamic changes in clustering or increases in long-range connections after stroke may be due to structural plasticity. Establishment or removal of synapses is required to achieve a homeostatic activity level.16 If tissue is damaged, remaining neurons will have a lower number of incoming connections. To meet a set-point of activity, new connections will be created locally as well as globally (long-range connections). Rewiring will impact network topology as it is crucial to retain small-world features.17 Preclinical animal studies are uniquely positioned to investigate these processes. For example, in nonhuman primates, tracer injections confirmed that a stroke in the primary motor region resulted in axonal sprouting from the adjacent premotor cortex.18 These new terminals and cell bodies appeared in the primary somatosensory region, supporting the idea that cortical areas distant from the injury undergo neuroanatomical reorganization. The increased long-range functional connectivity shown by this study could be a result of this process. The extent to which a brain can recover also depends on the capability of the network to functionally compensate for affected tissue. Degeneracy describes the ability of other brain regions to perform the same function as the lesioned regions.16,19 If a function can be retained, even though performance may be reduced, this is considered functional compensation. For example, optical imaging has shown that both adjacent (hindlimb motor) and distant (retrosplenial) regions assume function of an ischemic forelimb region.20 Compensation may be more difficult when there are large increases in functional connectivity and small-worldness.In summary, Blaschke et al1 2021 observed widespread increases in functional connectivity in both humans and mice, which highlights translational potential particularly when global or data-driven approaches are employed. The literature surrounding application of graph theory is diverse and while clustering may be vital for functional reorganization in rodents, increases in path length in humans may imply more long-range connections. An avenue for future research to establish whether stroke enhances interactions with distant regions could include observing functional connectivity during task performance. Because additional regions would be critical for improved integration, the increased functional connectivity and small-worldness should also occur during tasks. It would also be interesting to observe how the described acute changes are linked to long-term recovery or whether this is supported by structural connectivity changes. These can be detected by diffusion imaging and used to predict recovery or response to intervention.5,6,21,22 Finally, it would be useful to look at changes in individual regions or connections to better understand the connectome response network damage.Disclosures Dr Kaiser was supported by the UK Medical Research Council (MR/T004347/1). The other authors report no conflicts.FootnotesThe opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.For Disclosures, see page 2963.Correspondence to: Tracy D. Farr, PhD, School of Life Sciences, University of Nottingham, Medical School, Queen’s Medical Centre, Nottingham, NG7 2UH, United Kingdom. Email tracy.[email protected]ac.uk