Volume 4, Issue 4S_Part_3 p. T90-T91
Abstract
Free Access

IC-P3-211: Increased randomness of functional brain networks in AD: ‘Small-world’ network analysis of non-linear functional connectivity

Ernesto J. Sanz-Arigita

Corresponding Author

Ernesto J. Sanz-Arigita

VU Medical Center, Amsterdam, Netherlands

Search for more papers by this author
Menno M. Schoonheim

Menno M. Schoonheim

VU Medical Center, Amsterdam, Netherlands

Search for more papers by this author
Jeske S. Damoiseaux

Jeske S. Damoiseaux

VU Medical Center, Amsterdam, Netherlands

Search for more papers by this author
Serge A.R.B. Rombouts

Serge A.R.B. Rombouts

Leiden Institute for Brain and Cognition - LIBC, Leiden, Netherlands

Search for more papers by this author
Frederik Barkhof

Frederik Barkhof

VU Medical Center, Amsterdam, Netherlands

Search for more papers by this author
Philip Scheltens

Philip Scheltens

VU Medical Center, Amsterdam, Netherlands

Search for more papers by this author
Cornelis J. Stam

Cornelis J. Stam

VU Medical Center, Amsterdam, Netherlands

Search for more papers by this author
First published: 01 July 2008

Background

Alzheimer disease (AD) affects fundamental elements of the brain network dedicated to memory formation. However, its functional effects can spread to other brain systems through the so-called disconnection syndrome. Following this hypothesis, dysfunction and functional isolation of elements of a specific functional network will induce abnormal interactions between neuronal systems beyond the original network.

Methods

To test this hypothesis, 18 AD patients and 22 age-matched controls were scanned during resting state condition (e.g. lying awake with eyes closed) using FMRI (200 data-samples). 116 single time-series per subject were extracted by averaging all voxels’ time-series in non-overlapping regions of interest (ROI) defined in the anatomically labeled brain, AAL (1). Synchronization likelihood (SL) (2) was calculated as a linear and non-linear measure of the functional coupling between ROIs.

Results

We found increased synchronicity between frontal areas in contrast with a general decrease in posterior regions (Fig 1). To further describe the network revealed by the synchronization patterns, ‘small-world’ network parameters were computed (3, 4). Our results indicate significant group differences of ‘small-world’ descriptors; namely, the path length (e.g. average minimum number of steps between two network nodes) is significantly smaller in AD (mean path length AD=1.53; controls=1.69; t-test, P=0.012; values normalised by comparing to random networks).

Conclusions

This characteristic illustrates fundamental differences in the structure of the brain networks in AD, describing a more ‘random’ network type.

Details are in the caption following the image

Transverse section through the AAL template, showing individually labeled anatomical ROIs. Connecting lines represent significantly different SL values between AD and controls (p<0.05, uncorrected): solid lines – increased SL; dotted lines – decreased SL values.