Network graph analysis of category fluency testing.

Area of Science:

• Neuroscience
• Graph Theory
• Cognitive Science

Background:

• Category fluency, a measure of semantic memory, is often impaired early in Alzheimer disease (AD).
• Graph theory provides tools to analyze complex relationships within networks, examining properties like node connections and network structure.
• Key network properties include "small world properties" and "scale-free" characteristics, which describe network interconnectedness.

Purpose of the Study:

• To investigate the applicability of graph theory for analyzing category fluency data.
• To compare network graph properties across individuals with normal cognition, mild cognitive impairment (MCI), and AD.
• To characterize how network structures change with increasing cognitive impairment.

Main Methods:

• Category fluency task ('animals' in 60s) data from normal (n=38), MCI (n=33), and AD (n=40) participants were collected.
• Co-occurring item networks were constructed from fluency lists.
• Network variables, including clustering coefficients and path lengths, were analyzed and compared between groups.

Main Results:

• Small world properties decreased progressively from normal to MCI to AD groups.
• All groups exhibited scale-free network properties.
• Filtering low-connectivity nodes in normal and MCI networks produced graph characteristics similar to those observed in AD networks.

Conclusions:

• Network graph analysis offers a novel and promising approach for detecting and quantifying changes in category fluency associated with cognitive decline.
• The developed technique generates non-random network graphs with properties consistent with established graph theory models.
• These findings suggest graph theory can effectively differentiate cognitive states based on semantic network structures.