Effects of Neonatal Sound Deprivation in Animal Research

Elsewhere in this issue are presented complete literature reviews that are consistent with, but do not prove, the hypothesis that the mild-to-moderate hearing losses associated with recurrent or chronic otitis media in young children can cause communicative disorders. The hypothesis has not been, and perhaps cannot be, proved, since lack of rigor is inherent in long-term retrospective human studies (Hanson and Ulvestad, 1979; Ventry, 1980; Paradise, 1981.) There have been several nonhuman studies, however, that are relevant to this hypothesis; this article will review them. Although most were undertaken to explore the basic mechanisms of the development of hearing, they may also help answer the important clinical question: Is there a causal relationship between postnatal conductive hearing loss and auditory behavior, physiology, and/or morphology? At the outset we must recognize that there are both advantages and disadvantages in using nonhuman experiments to answer questions about human problems. The nonhuman studies we are interested in have usually been performed on rodents, with the conductive losses imposed either by removing the auditory ossicles or by blocking the external auditory meatus, both of which result in a conductive loss of 40 to 55 dB; this is considerably greater than the 15 to 30 dB loss suffered by children with otitis media. Furthermore, these conductive losses have been of constant duration in the rodents, whereas in otitis media children they are usually fluctuating. Moreover, hearing does not begin in most rodents until 8 to 12 days after birth and is not mature until 16 to 24 days, whereas newborn human infants already have mature hearing (although not mature processing). Finally, since we are concerned with speech and language, which are unique to the human, we must also be concerned with whether the central mechanisms of speech processing are unique to the human; if so, it may be inappropriate to apply data from experimental animals to this particular human problem. On the plus side, nonhuman experimentation can be rigorously controlled, whereas human studies cannot be. Conductive losses can be created at known times, independent of any pathology. One can design experimental and control groups. By using inbred strains, one can even approach genetic homogeneity. Diet, socioeconomic background, and medical history are not a problem, and the subjects do not move out of town before the experi-