JOURNALOF NEUROPHYSIOLOGY Vol. 67, No. 1, January 1992. Printed in U.S.A. Encoding of Sound-Source Location and Movement: Activity of Single Neurons and Interactions Between Adjacent Neurons in the Monkey Auditory Cortex MERAV AHISSAR, EHUD AHISSAR, HAGAI BERGMAN, AND EILON VAADIA Department of Physiology, Hadassah Medical School, The Hebrew University, Jerusalem 91010, Israel SUMMARY AND CONCLUSIONS 1. Neuronal mechanisms for decodingsoundazimuth and an- gular movementwerestudied by recordings of several single units in parallelin the core areas of the auditory cortex of the macaque monkey. The activity of 188units was recorded during the presen- tation of moving and static sound stimuli. Both the activity of single units and the interactions betweenneighboring neuronsin response to eachstimulus were analyzed. 2. Sixty-two percent of the units showedsignificant modula- tion of their firing rates as a function of the stimulus azimuth. Contralateral stimuli were preferred by the majority (--60%) of theseneurons.Thirty-five percent of the units showed mild but statistically significantmodulation of their firing rates,which was specifically attributed to the angular movement of the sound source. 3. Eighty-nine percentof the “movement-sensitive” units were also “azimuth sensitive.”The sound source’s azimuth determined the pattern of the response components (on, sustained, off), whereas the source’s movement affected only the magnitude of these components, typically the sustained component. Most neu- rons for which the sustained response to static sounds was greater for contralateral than ipsilateralstimuli preferred moving sounds that weremoving into the contralateral hemifield. 4. Cross-correlation analysis was carried out for 245 neuron pairs. Cross-correlograms were computed for each pair under all stimulus conditionsto allow comparison of the neuronal interac- tions under the various conditions. The shapes of some correlo- grams (after subtraction of direct stimulus effects) weredependent on specific stimulusconditions, suggesting that the effective con- nectivity between these neurons depended on the location and/or movement of the sound stimuli. Furthermore, joint peristimulus time (JPST) analysis indicated that modifications of connectivity may be temporally related to the stimulus and may occur over shortperiods of time. These results could not have beenpredicted from analysis of the independent single-unit responses to the stimuli. 5. The data suggest that both firing rates and correlatedactivity betweenadjacent neuronsin the auditory cortex encode sound location and movement. INTRODUCTION The sensitivity of single units in the auditory cortex to the spatial location of sound has been studied by several re- searchers (Benson et al. 198 1; Eisenman 1974; Imig et al. 1990; Middlebrooks and Pettigrew 198 1; Rajan et al. 1990). They found that more than one-half of the studied units were spatially tuned to some extent. Most of these units showed preference for stimuli presented from the contralat- era1 hemifield. Ablation studies indicated that the primary auditory cortex (AI) is essential for localization of short- duration sounds in the contralateral hemifield (Jenkins and Masterton 1982; Jenkins and Merzenich 1984). Jenkins and Merzenich selectively ablated AI or all other auditory cortical areas of cats and showed that AI is necessary and sufficient for the performance of a sound localization task. The spatial tuning curves of the azimuth-sensitive units are relatively wide all along the auditory pathways. In fact, even the neurons with the narrowest spatial tuning curves are less precise than the monkey’s behavioral acuity in local- izing a sound source that is -3-4O (Brown et al. 1978). This suggests that localization is coded by a population of neurons, rather than by single neurons. Indeed, Eisenman has suggested that acuity is achieved by the activity in groups of widely tuned units, implying that connectivity among neurons plays an essential role in encoding stimulus location. Nevertheless, the connectivity within neuronal groups has hardly been examined because of the technical difficulty of a simultaneous recording of the activity of sev- eral neurons. Bloom (1984) and Bloom and Gerstein (1986) studied the effect of binaural stimuli on the activity of single units in the auditory cortex of cats by the use of simultaneous recording of several neurons. He found that functional interactions between neuron pairs could depend on binaural parameters. An appealing suggestion for an additional role of neurons in the auditory cortex is that of coding of sound movement. Location information may serve as a substrate for move- ment detection, and movement detection may affect dis- crimination of sound-source location. Therefore it may be advantageous for the nervous system to compute both loca- tion and motion within the same area. Indeed, a study by Altman and Kalmykova (1986) indicated that bilateral le- sions of the auditory cortex of dogs impaired the animals’ ability to detect simulated sound-source movement. The hypothesis that neurons in the cat auditory cortex are involved in detecting movements of sound has been tested by Sovijarvy and Hyvarinen (1974). They applied (manually driven) moving acoustic stimuli and reported a small number of neurons that were specifically sensitive to movements of the sound source. This line of research was not followed by further studies despite the encouraging re- sults. The present study was designed to further investigate the sensitivity of cortical neuronal activity to azimuth and angu- 0022-3077192 $2.00 Copyright 0 1992 The American Physiological Society 203