296
M. Thanaj et al.
to define the spectral bands as originally described by Stefanovska and colleagues
[8]. It is unlikely that the boundaries of these frequency intervals remain constant
across a cohort of individuals, or for a given individual, under changing conditions
of physiological stress. State-dependent fluctuations in frequency intervals may thus
give rise to different spectral signatures within, and across, the cohorts studied.
However, frequency domain analysis remains a valuable tool in our understanding of
the processes modulating microvascular BF and their relative contribution to overall
network perfusion [55].
19.4 Information and Complexity-Based Analysis
of Microvascular Blood Flow Signals
While conventional time and frequency domain analysis techniques have proved
valuable in the understanding of blood flow within a microvascular bed they have
failed, so far, to describe mechanistically the changes in observed flow patterns
between pathological conditions or haemodynamic states. Nonlinear methods, based
on ideas from information theory, have been used to quantify the regularity and
randomness of short lengths of physiological signals and have demonstrated the
potential for diagnostic capability [6]. Recently, their application has been extended
to the LDF signal from superficial microvascular networks.
Complexity analysis quantifies the degree of variability or loss of spontaneity in
a time series and has been applied to a range of bio-signals, including electroencephalograms [43] and electrocardiograms [86]. The degree of variability in these
signals reflects the physiological adaptability of the underlying system and is an
established biomarker of overall health status [1]. There is, however, no single definition of complexity. Nagaraj and Balasubramaian [58] describe three methods of
quantifying complexity which in the context of the current discussion relate to (i)
how hard it is describe the information in the LDF signal, i.e. the number of unique
patterns in the time series or model order, (ii) how hard it is to create or loosely
compress the information and (iii) the degree of organisation or structure in the
LDF signal. The most widely used measures in LDF signal analysis are Lempel-Ziv
Complexity (LZC), sample entropy and effort-to-compress (ETC) complexity. All
these approaches have successfully provided a measure of the information content
of the LDF signal [83, 85] (Fig. 19.1).
19.4.1 Lempel-Ziv Complexity (LZC)-Based Analysis
Lempel-Ziv complexity (LZC) [50] provides a measure of how difficult it is to
describe the information contained in a signal and is the length of the shortest instruction set needed to reconstruct the signal without information loss. A simple periodic
M. Thanaj et al.
to define the spectral bands as originally described by Stefanovska and colleagues
[8]. It is unlikely that the boundaries of these frequency intervals remain constant
across a cohort of individuals, or for a given individual, under changing conditions
of physiological stress. State-dependent fluctuations in frequency intervals may thus
give rise to different spectral signatures within, and across, the cohorts studied.
However, frequency domain analysis remains a valuable tool in our understanding of
the processes modulating microvascular BF and their relative contribution to overall
network perfusion [55].
19.4 Information and Complexity-Based Analysis
of Microvascular Blood Flow Signals
While conventional time and frequency domain analysis techniques have proved
valuable in the understanding of blood flow within a microvascular bed they have
failed, so far, to describe mechanistically the changes in observed flow patterns
between pathological conditions or haemodynamic states. Nonlinear methods, based
on ideas from information theory, have been used to quantify the regularity and
randomness of short lengths of physiological signals and have demonstrated the
potential for diagnostic capability [6]. Recently, their application has been extended
to the LDF signal from superficial microvascular networks.
Complexity analysis quantifies the degree of variability or loss of spontaneity in
a time series and has been applied to a range of bio-signals, including electroencephalograms [43] and electrocardiograms [86]. The degree of variability in these
signals reflects the physiological adaptability of the underlying system and is an
established biomarker of overall health status [1]. There is, however, no single definition of complexity. Nagaraj and Balasubramaian [58] describe three methods of
quantifying complexity which in the context of the current discussion relate to (i)
how hard it is describe the information in the LDF signal, i.e. the number of unique
patterns in the time series or model order, (ii) how hard it is to create or loosely
compress the information and (iii) the degree of organisation or structure in the
LDF signal. The most widely used measures in LDF signal analysis are Lempel-Ziv
Complexity (LZC), sample entropy and effort-to-compress (ETC) complexity. All
these approaches have successfully provided a measure of the information content
of the LDF signal [83, 85] (Fig. 19.1).
19.4.1 Lempel-Ziv Complexity (LZC)-Based Analysis
Lempel-Ziv complexity (LZC) [50] provides a measure of how difficult it is to
describe the information contained in a signal and is the length of the shortest instruction set needed to reconstruct the signal without information loss. A simple periodic