Patterning Brain Developmental Events via the Discovery of Time-Series Coherences
George Potamias, Catherine R. Dermon, Vassilika Vouton · 2001
ABSTRACT: In this paper we present a methodology concerned with data-mining issues on neuro-physiological Time-Series data. The data refer to local cerebral biosynthetic activity during late embryonic development. The methodology is realised by the introduction of a novel algorithmic process and related formulas for the discovery of coherences in time-series data. The inovation comes from the inclusion of specific ‘control ’ operations in the elaborated time-series matching metric. The final outcome is the clustering of time-series into similar-groups. Clustering is performed via the appropriate customization of a phylogeny-based clustering algorithm and tool. Then, with a close and careful examination of the formed phylogenetic cluster-trees we are able to reveal indicative brain developmental patterns. The whole system (i.e., respective classes for the time-series similarity/distance computation and the phylogeny-based clustering) is built in Java. Experimental results and their neuro-physiological interpretation shows the reliability and efficiency of our approach.