%0 Generic
%A Schmidt, Jan-Philip
%C Heidelberg, Germany
%D 2012
%F heidok:14758
%R 10.11588/heidok.00014758
%T An analysis of age-related alterations in functional memory networks
%U https://archiv.ub.uni-heidelberg.de/volltextserver/14758/
%X The human brain is the most complex organ of the human body and many aspects of its functioning have not yet been understood. One of the most fascinating abilities of the human brain is the skill to store and retrieve information, which is what we refer to as memory. One attempt to get a deeper insight into the functioning of memory is to analyze the complex activity pattern of the human brain that emerges while a memory task is being processed. The understanding of memory is epistemologically very intriguing since it is this ability that enables us to collect, to store and to recall ideas, emotions and thoughts - hence, it builds our own identity. This thesis analyses age-related changes in functional connectivity networks related to episodic and working memory processing. The data for this study were measured using fMRI technique and the sample set consisted of healthy individuals aging from 20 up to over 80 years. Based on the fMRI data we construct correlation networks by correlating pairwisely the measured voxel activity, the nodes of the network being brain voxels, the edges being  correlations. These networks are thresholded, anatomically clustered and analyzed by computing statistical network measures, using spectral methods, computing network entropy and calculating persistent homology. The main findings are: elderly individuals exhibit expanded neural networks with less differentiation between episodic and working memory tasks. However, we observe compensatory mechanisms that accompany this dedifferentiation process. Network synchronizability is higher for elderly individuals. Network entropy increases as well with age, yielding a lower network vulnerability  for elderly individuals. Aging processes leave traces in the homology pattern of the networks, whereas all brain networks exhibit different persistent homology features.