Recent advances in the care of premature infants have led to increasing attempts at sensory-motor intervention in order to enhance future potential functioning of the infants. Concommittant with such programs of increased intervention, there arises a need for appropriate instruments to objectively evaluate infant behavior and thereby to assess the efficacy of these interventions. Most previous studies of infant temperament have focused upon parent questionnaires administered after discharge. However, instruments to objectively assess behavior in these babies while still in the NICU are currently non-existent. Irritability is one component of behavior which has been shown to place the premature at high risk. This study was designed to assess neonatal irritability during hospitalization and to develop a tool to objectively score infant irritability. Methods: Our study utilized observations self regulatory behavior patterns, focusing on 11 approach and 10 withdrawal reactions. The design was based on the non-interventional observation components of Als' APIB (Assessment of Preterm Infant Behavior). Infants were observed for 20 minutes after their clinical condition had stabilized. An irritability score, defined as withdrawal behavior/(withdrawal+ approach) behavior, was assigned for each observation period. Results: The instrument was performed on 42 premature and 15 full term infants. Mean irritability profile was strongly correlated with gestational age: prematures - 30.6 vs. full term- 9.8; p<0.0001. There was no influence of gender. Speculation:We speculate that the development of an objective tool to evaluate irritability will enable us to begin to identify infants at highest risk for future problems and thereby those most likely to benefit from infant stimulation. Further studies are needed to ascertain whether such an irritability profile can be correlated with other clinical risk factors.