International Workshop on Models and Analysis of Vocal Emissions for Biomedical Applications (MAVEBA 1999)

Florence, Italy
September 1-3, 1999

A Non-Linear Approach to the Parametric Analysis of Spontaneous and Pain Cry in the Human Infant

A. Accardo (1), A. Clarici (2), U. de Vonderveidt (3), M. Caselli (2), A. Garau (1), C. Travan (2)

(1) D.E.E.L; (2) Department of Physiology and Pathology; (3) Department of Neonatology, Paediatric clinic. Children's Hospital "Burlo Garofolo"
University of Trieste, Italy

In order to investigate whether nonlinear methods of signal analysis provide a useful quantification for a successive classification of newborn cries, the present study computed the fractal dimension (D), the standard deviation (SD), the zero crossing, the fundamental frequency (F0) and the formant frequencies (F1 and F2) on cry segments produced by newborns in spontaneous way or during pain. The results demonstrated that a suitable algorithm, mainly based on SD, D and F0, can be used to identify the characteristic pattern of a single cry event. A combination of these parameters allowed us to correctly classify the cries, demonstrating that linear and non linear parameters can be successfully jointly used in the cry analysis.

Full Paper

Bibliographic reference.  Accardo, A. / Clarici, A. / Vonderveidt, U. de / Caselli, M. / Garau, A. / Travan, C. (1999): "A non-linear approach to the parametric analysis of spontaneous and pain cry in the human infant", In MAVEBA-1999, 103-107.