G. Dwyer

EMG burst waveform recognition procedure

A fast, simple EMG (electromyogram) burst waveform recognition algorithm has been developed for a personal computer. Raw EMG data are detrended, squared, and filtered. A sample EMG waveform segment, known as a template, is drawn from this data. A recognition signal is constructed as the convolution of the template and the prepared EMG data. This recognition signal is interpreted to yield the position, duration, and strength of individual EMG bursts.<<ETX>>

A distributed microcomputer-controlled system for data acquisition and power spectral analysis of EEG

A relatively powerful and inexpensive microcomputer-based system for the spectral analysis of the EEG is presented. High resolution and speed is achieved with the use of recently available large-scale integrated circuit technology with enhanced functionality (INTEL Math co-processors 8087) which can perform transcendental functions rapidly. The versatility of the system is achieved with a hardware organization that has distributed data acquisition capability performed by the use of a microprocessor-based analog to digital converter with large resident memory (Cyborg ISAAC-2000). Compiled BASIC programs and assembly language subroutines perform on-line or off-line the fast Fourier transform and spectral analysis of the EEG which is stored as soft as well as hard copy. Some results obtained from test application of the entire system in animal studies are presented.

EKG waveform recognition procedure

A fast, accurate, and robust EKG (electrocardiogram) waveform-recognition algorithm has been developed for a personal computer. The method uses the convolution of raw EKG data with one or more sample EKG waveform templates as a recognition signal. Recognition occurs when the convolution exceeds a user-defined reference level. This level can depend on the history of the recognition signal as well as on other physiological data.<<ETX>>

Inverse power-law distributions and 1/F power spectra in fetal breathing dynamics

It is shown that fetal breathing dynamics demonstrate an inverse power-law relationship in the distributions of the inter-burst intervals and in the power spectrum. Furthermore, it is shown how these fractal characteristics change as a function of gestational age of the fetus. The observation of the inverse power-law relationship in the fetal breathing dynamics supports the proposal that fractal mechanisms may be involved in the regulation of breathing rate variability, giving rise to self-similar fluctuations over multiple time scales.<<ETX>>

Breathing modulates fetal heart rate

To assess the modulation of fetal heart rate (FHR) by breathing movements, the authors investigated the relationship between FHR and fetal breathing rate (FBR). FHR was obtained from the R-R interval of the EKG and FBR from the burst-burst interval of the diaphragmatic electromyogram (EMG). The R-R intervals of EKG and burst-burst interval of the diaphragmatic EMG were determined using a template recognition method. These data were then analyzed in both the frequency and time domains. Frequency analysis was carried out using the linear prediction method with the optimum order. It is confirmed that fetal breathing movements modulate fetal heart-rate variability.<<ETX>>

1/f Fluctuations of breathing rate in fetal sheep and IPFM model

Fetal breathing rates (FBR) in sheep have been investigated using spectral analysis techniques. The power spectral densities of the FBR were found to be inversely proportional to Fourier frequency (1/f spectrum) below 0.06 Hz. The levels of the 1/f spectrum are same with changing the morphine dose, when normalized by the mean square of the data. In order to simulate 1/f fluctuation of the FBR, the integral pulse frequency modulation (IPFM) model is used, and parameters were fluctuated various random noise. As a result, 1/f fluctuation similar to the experimental data is obtained with the 1/f fluctuation of the parameters.

Change Of Fractal Dimensions Of Breathing Rate In Fetal Sheep

We found already that the behavior of fetal breathing movements were not random by way of the l/f fluctuations, clustering and selfsimilarity of breathing rate in fetal lambs chronically instrumented intrauterine. In order to investigate the dynamic behavior of the fetal breathing movements more quantitatively, we calculated correlation dimensions and determined the fractal dimensions with increasing embedding dimensions. The fractal dimension was non-integer and saturable, which meant the fetal breathing movements was different from random noise. Furthermore, the correlation dimensions of fetal breathing rates changed with maturation of fetal sheep in uterine.

Analysis Of Dynamic Behavior Of Breathing Rate In Fetal Sheep Using Fractal Dimensions

We have previously found that fetal breathig movements (FBM) do not occur in a random manner, and are characterized by properties of self-similarity, l/f fluctuations and clustering. In an attempt to further characterize the dynamic behavior of FBM, we have now calculated correlation dimensions and determined the fractal dimension with increasing embedding dimensions. The fractal dimension was found to be non-integer and saturable, thus demonstrating that the dynamics of FBM is different from random noise.

Spectral analysis of heart rate variability using diaphragmatic EMG burst activity as a time base

Spectral analysis of heart-rate variability (HRV) normally reveals a harmonic mode associated with breathing activity. Since this mode is locked to the breathing rate, temporal variations in breathing rate lead to a broadening of the HRV spectral peak width. It is shown that specific knowledge of breathing activity can be used to eliminate this broadening.<<ETX>>

Spectral analysis of heart rate variability using breathing bursts as a time base

Diaphragmatic EMG burst activity is analyzed by means of waveform recognition, yielding a nonlinear time scale where the interval between each EMG burst is taken as one time period. The heart-rate variability data are then suitably interpolated to yield a pseudotime data series with respect to the EMG time base. Spectral analysis of the interpolated EKG rate data reveals a harmonic mode associated with breathing, but with substantially reduced peak width. Using this procedure, the EKG spectrum is averaged over long periods with little peak broadening due to changes in breathing rate.<<ETX>>