IMPORTANCE OF NMDA RECEPTORS FOR MULTIMODAL INTEGRATION IN THE DEEP LAYERS OF THE CAT SUPERIOR COLLICULUS

1. Many sensory events contain multimodal information, yet most sensor y nuclei are devoted to the analysis of single-modality information. I n the deep superior colliculus (DSC), visual, auditory, and somatosens ory information converges on individual multimodal neurons. The respon ses of multimodal neurons are determined by the temporal and spatial c orrespondence properties of the converging inputs such that stimuli ar ising from the same event elicit a facilitated multimodal response. 2. N-methyl-D-aspartate (NMDA) receptors may underlie the detection of s patial and temporal coincidence and could be involved in the generatio n of multimodal facilitatory responses because of the nonlinear proper ties of NMDA-receptor-mediated events. To assess the role of NMDA rece ptors in multimodal integration, we made extracellular recordings from single multisensory neurons in the DSC of the cat. 3. The responses t o visual, auditory, and somatosensory stimuli alone and to multimodal combinations of stimuli were challenged with iontophoretically applied D-2-amino-5-phosphonovalerate (AP5), an NMDA receptor antagonist. All responses to visual stimuli presented alone (n = 9) were greatly redu ced. Somatosensory responses (n = 25) were usually decreased. In contr ast, the responses to auditory stimulation were decreased (n = 9), una ffected (n = 3), or enhanced (n = 5). 4. Responses to multimodal stimu lus presentations were consistently reduced during iontophoretic appli cation of AP5, irrespective of the modalities that made up the stimulu s. The reductions of multimodal responses were significantly greater t han the sum of the reductions of responses to single-modality stimuli. 5. The data suggest that for unimodal stimuli, the importance of NMDA receptors in synaptic transmission of sensory responses in DSC may be dependent on the stimulus modality. Furthermore, NMDA receptors are of major importance in the integration of input from different modalitie s for the generation of multimodal responses.