Perhaps the most important finding from the IRAS survey was the discovery of a significant population of galaxies whose infrared luminosities are equivalent to the bolometric luminosities of optically-selected quasars. Assessing the relative importance of an AGN and massive stars for powering ultraluminous IR galaxies (ULIGs) is essential for understanding these objects as massive galaxies and/or quasars in formation, and for testing suggestions of an evolutionary connection between starburst and AGN phenomena. The target galaxies were taken from a recently compiled sample of ULIGs which contains the brightest objects at a given L_ir (< 10^13 L_sun) and which is not biased towards `warm' quasar-like objects with large S(25 um)/S(60 um) ratios. The galaxies in our sample span the luminosity range log[L_ir/L_sun] = 12.3 - 13, currently not explored spectroscopically by the ISO Central Program. The Central Programme focuses on previously published objects that typically lie below this range of L_ir, and the few exotic `hyperluminous' (L_ir > 10^13 L_sun) objects at much higher redshifts. Our sample therefore provides an important bridge between these two extremes that could prove critical to the interpretation of the more distant `hyperluminous' objects. We plan to measure the fluxes of the fine-structure lines [Ne II] 12.8 um, [Ne III] 15.6 um, [Ne III] 36.0 um, [Ne V] 14.3 um, [O III] 52 um, and [O III] 88 um. These data will then be combined with our optical and near-infrared measurements to (1) determine the physical conditions and ionization mechanisms of the nebular gas in these galaxies, (2) measure the stellar content in the objects powered by stars, and (3) evaluate the importance of dust extinction. This study will provide new insights into the relative role of starbursts and AGN in powering ULIGs, and possible evolutionary links between these two phenomona.