Authors: [tex2html_wrap4248]M. Petroni, C.C. Johnston (School of Nursing, McGill University), A.S. Malowany
Investigator username: malowany
Subcategory: language and program understanding
The human infant depends heavily on vocalizations for indicating its state to care-givers, be they parents or health-care professionals. Studies have shown that infants who are at ``high risk'', due to events such as complicated birth histories for example, have cries which are more arousing and urgent sounding than those of other infants. It is therefore of interest to identify cry features associated with certain pathological conditions, thus potentially expanding the possibilities for medical diagnosis using a readily available and non-invasive parameter, such as the cry. Neurological considerations motivate an analysis on the vocal fundamental frequency [tex2html_wrap4246] and artificial neural networks (ANNs) are being trained to differentiate between so-called normal cries and pain cries, or the cries of infants who are labelled as being ``at risk''.