Measuring Infant Pain Objectively using Sensor Fusion and Machine Learning Algorithms

Newborns are routinely and frequently exposed to pain during Neonatal ICUNICUcarePain assessments in
neonates are difficultlabor intensivesubjective and unreliableoften resulting in excessive or inadequate
analgesiaOur overall objective is to measure infant pain objectivelyreliablyand in real timeWe will extract
pain related information from multiple non invasive sensorsdevelop a sensor fusion framework to integrate
multi modal sensor data into a single pain scoreand assess the validity of this approach by comparing with
validated clinical pain scoresSpecific aimsTo differentiate acute pain from baseline or non painful
eventswe will studynewborns usingfacial electromyographyEMGto record facial expressions specific
for infant painelectrocardiographyECGto measure heart rate changes and heart rate variabilityskin
conductance to measure catecholamine dependent palmar sweatingelectroencephalographyEEGusingactiveelectrodes to assess pain related brain activityand pulse oximetrySpOto record pain induced
changes in oxygenation and peripheral perfusionWe will study acute painful procedures associated with mildmoderateor severe pain inlate pretermweeksandterm newbornsweeksBedside
nurses will use validated pain scoring methods to concurrently assess these infants for painA pain expert will
independently assessof subjectsto establish inter rater reliability and to authenticate the bedside nursespain scoresFrom each sensorwe will extract pain related data that correlate strongly with the clinically
relevant pain scoresTo develop sensor fusion frameworks integrating data from multiple sensorsProprietary machine learning algorithms will fuse pain related data from allsensorscalibrateitself for each
newborn by using data from prior pain eventsand compensate for missing or unreliable dataSensor fusion
frameworks including combinations of these sensors will help to identify infant pain with far greater specificity
and sensitivity than the subjective pain scales used clinicallyProcedures will be included to assess the scaling
properties of this objective approach and to refine the principal algorithmsData analyses will assess inter rater
reliability and internal consistencyverify contentconcurrent and construct validityand include multivariable
modeling for optimal selection and weighting of the sensor variables that will compute the final objective pain
scoreThis approach will eventually lead to a bedside ICU monitorcompatible with the ECGSpOEEGEMGand skin conductivity sensorswhich displays the current pain intensity and trends within the time
periods of clinical interestAn objectiveautomated pain detection device developed for newbornsand
adapted for other nonverbal patientswill reduce the subjectivity and variability of pain assessmentsimprove
the safety and efficacy of various analgesics used for treating neonatal painavoid the acute side effects and
long term effects of both unrelieved pain or excessive analgesia in newbornsprevent iatrogenic tolerance and
neonatal abstinence syndromereduce the workload of bedside NICU nurses and improve clinical outcomes!Newborns receiving intensive care in the Neonatal ICUNICUare repeatedly exposed to acute painful
procedures during routine medical carebut it is difficult to determine if they are experiencing pain or notor
their response to pain relieving therapiesIn this pilot studywe will use a novel machine learning framework to
develop an automated bedside monitor that is designed to measure pain intensity in newborn infantsobjectivelyreliablyand in real timecapable of displaying the current pain score as well as trends within time
periods of interest to bedside clinicians or parentsReliably measuring pain in newborns will enhance the
safety and efficacy of pain relieving drugslike morphinefor treating pain in newbornsthus avoiding the
immediate side effects as well as the long term detrimental effects from unrelieved painversus excessive or
highly variable drug therapy in the newborn period