The Project

Various pathological conditions can arise during the first year of life of an infant, which calls for immediate detection and intervention, considering that parents cannot always identify the signs of pathology that specific movements and sounds of the baby imply. For instance, although seizures with intense convulsions can be easily identified, there are certain epileptic syndromes that are manifested only through specific body postures and face movements. Such subtle signs of epilepsy can be easily ignored by inexperienced observers. Furthermore, many pathological situations can occur during the infant night sleep and potentially harm the baby if not detected promptly: arrhythmias, vomiting, sudden infant death syndrome, breathing disorders and apnea, or epileptic and febrile convulsions.

Nevertheless, the imperative need for early detection of medical emergencies during infant night sleep should not cancel the significance of an unobtrusive and non-invasive detection system. Invasive devices and sensors could disrupt the extremely important for their development sleeping phases and degrade the quality of babies’ rest.

The main objective of the xVLESPIS project is the development of an integrated and non-invasive biosignal recording system from the baby’s cradle, that will identify suitable biomarkers for the detection of potentially hazardous pathological conditions. Technavio’s analysts forecast the global baby monitors market to grow at a CAGR of 10.92% during the period 2017-2021. Parents’ need for increased baby safety actually drives the evolution of this market. The proposed product is expected to be a significant addition to the current market, adding clinically meaningful alarming and recommendation capabilities, something that is currently missing from the baby monitor market.

The system to be developed by the team of xVLEPSIS will entangle diverse user-friendly electronic smart sensors, integrated under a ‘smart’ mattress, in combination with a high resolution baby monitor. In brief, the following biosignals will be recorded, analyzed and investigated for their applicability as biomarkers for certain pathologies:

• Video recording using a high resolution camera and audio recording using a microphone of high definition.
• Ballistocardiogram (BCG) recording, which is a technique for producing a graphical representation of repetitive motions of the human body arising from the sudden ejection of blood into the great vessels with each heart beat and it can be recorded using pressure sensors under the bed mat.
• Temperature and humidity detection using suitable sensors under the bed mat.

The development of an intelligent system that will detect potentially hazardous pathological conditions, with the use of sophisticated machine learning techniques will lead to:

1. a mobile or smart watch based notification system, that will alert the parents in the case of emergency, and
2. the continuous biosignal recording, throughout the infant night sleep. The recorded data could be sent to the doctor or the hospital, in the case an abnormality is detected, or they could be evaluated by the doctor during regular infant examination, in the case nothing critical is detected. Therefore, the pediatrician will be able to examine and evaluate all the available medical data and detect any incidents that may have occurred at night and have not been perceived by the parents.

Many advantages arise from the development of a low cost product with all the aforementioned features:

• Continuous recording of high definition video and audio will allow for a more effective monitoring of the infant, whereas the pathological situations detection system will lead to the discovery of incidents that would otherwise remain unnoticed.
• The proposed non-invasive monitoring system will aid the diagnosis and proper treatment of medical disorders that can occur while the parents are not present, e.g. febrile convulsions, epileptic seizures or apnea.
• The pediatricians always face the challenge to evaluate medical incidents solely based on the information that parents provide, which is not objective and accurate, especially during the first year of life of the infants. The proposed integrated system offers the medical professionals in charge the opportunity to assess those incidents based on detailed recorded biosignals and, thus, form a better opinion on the diagnosis.
• The use of innovative machine learning algorithms performed on the multimodal medical signals will significantly aid the detection of new quantitative biomarkers of the relevant diseases.
• The medical database that will be implemented will significantly contribute to the research and study of such early childhood disorders.