Functional network disruption in the degenerative dementias

Abstract

Despite advances towards understanding the molecular pathophysiology of the neurodegenerative dementias, the mechanisms linking molecular changes to neuropathology and neuropathological changes to clinical symptoms remain largely obscure. Connectivity is a distinctive feature of the brain and the integrity of functional network dynamics is crucial for normal functioning. A better understanding of network disruption in the neurodegenerative dementias might help bridge the gap between molecular changes, pathological changes, and symptoms. Recent findings on functional network disruption as assessed with resting-state or intrinsic connectivity functional MRI and electroencephalography and magnetoencephalography have shown distinct patterns of network disruption across the major neurodegenerative diseases. These network abnormalities are somewhat specific to the clinical syndromes and, in Alzheimer's disease and frontotemporal dementia, network disruption tracks the pattern of pathological changes. These findings might have practical implications for diagnostic accuracy, allowing earlier detection of neurodegenerative diseases even at the presymptomatic stage, and tracking of disease progression.

Figure 1

The pathophysiological framework of neurodegenerative diseases: connectivity as an intermediate phenotype between pathology and symptoms. The case of AD.

Figure 2

Functional connectivity on resting-state fMRI in healthy subjects. ICA-derived resting-state fMRI networks (DMN, salience, left and right executive-control, visual and motor networks)^- of a healthy 33-year old male. Red-to-yellow colours indicate the strength of each voxel's connectivity to overall component time series (shown beneath each map).

Figure 3

Functional connectivity of resting-state EEG/MEG in healthy subjects. Headplot showing functional MEG network of a healthy 63-year old female in the alpha (8-13 Hz; left) and beta (13-30 Hz; right) frequency ranges. Coloured lines indicate different functional sub-networks (modules), black lines represent their interconnections (only visualized in beta band example). Background colours indicate connectivity strength (red indicates hub – i.e. highly connected -regions). SL=synchronization likelihood. A=Anterior; P=Posterior; L=Left; R=Right.

Figure 4

Structural connectivity assessed with DTI in a healthy (33-year old male) subject. DTI-tractography identifies long (mainly visible in sagittal view as green and blue colour-coded fibers) and short (mainly visible in axial and coronal views as red colour-coded fibers) WM connections. Specific tracts can be identified which subserve distinct cognitive and non-cognitive functions. The fornix and cingulum are mainly associated with memory and emotional processing, cortico-cortical association and intra-hemispheric tracts are associated with a broad range of cognitive processes, the corticospinal/cerebellar tracts are generally involved in motor disorders.

Figure 5

Schematic representation of (B) ‘small-world’ brain functional network and of (A) simulated ‘regular’ and (C) ‘random’ networks with the same number of nodes (n=35) and connections (n=120). (A) Regular networks have many connections among neighbouring regions (red lines) and few connections with distant nodes (light blue lines). (B) Small-world networks have less local connections and more long distance connections. (C) Random networks have few local connections and many connections among distant regions. Each network is shown overlaid onto a standard template (upper row) and in schematic representation (middle row). Nodes represent 35 cortical points of the left hemisphere drawn from the Automated Anatomical Labeling template, and edges represent functionally connected nodes. The real-world network was extracted from a single subject, the corresponding regular (A) and random (C) networks were simulated using the Brain Connectivity Toolbox. The corresponding theoretical Watts-Strogatz network models are also shown (lower row; adapted from ref ). Reproduced from Nature Publishing Group (permission requested).