Information theoretic covariance complexity and its relation to pattern recognition

The labile brain. II. Transients, complexity and selection

“…The former approaches are based on some measure of the entropy of the system (e.g. Morgera 1985) and can be related to algorithmic complexity (framed in terms of the minimum length of an algorithm required to generate an observed time-series). More recently, entropy-based complexity measures have been proposed that try to capture the balance between integration among di¡erent neuronal systems and the preservation of information that is unique to them.…”

On the meaning and interpretation of global descriptors of brain electrical activity. Including a reply to X. Pei et al.

“…There the partitions of unity of interest are probability distributions, while in the present context function H is applied to distributions of signal power (variance) over spatial modes; therefore it cannot be interpreted as an information entropy. Nonetheless, a connection is given by the work of Morgera (1985), who derived the related concept of covariance complexity via information theoretic considerations. A generalised version of spatial complexity, (Rényi, 1961) that includes Shannon's entropy as the limiting case for α→1.…”

Correlation Statistics for cDNA Microarray Image Analysis

“…In the present study, we investigated the correlation aspects of pixels comprising a spot using Morgera's covariance complexity [11]; subsequently, two measures of correlation, namely, Pearson's correlation (parametric) and Spearman rank correlation (nonparametric) [12] are proposed to determine the foreground and background intensity of the given spot. These statistics are also used to flag poorly hybridized spots, thus minimizing falsepositives.…”

“…Subsequently, we used the Morgera's covariance complexity measure () [11] to quantify the correlations of the pixels comprising spots A, B, and C, hence addressing question QA. A description of the computational procedure of () is shown in Fig.…”