Cerebral representation of vocalization in the squirrel monkey.

Abstract

Specific vocalization types following electrical stimulation of 5940 electrode positions are studied in 39 squirrel monkeys. Except cerebellum, caudal medulla, and a few cortical areas, the sites of stimulation were distributed throughout the brain. Each vocalization elicited was tested for reproducibility at the site of stimulation and in homologue structures. All vocalizations were analyzed spectrographically and then classified for eight fundamental vocalization types. The cerebral distribution of two call types forms continuous systems extending from midbrain via diencephalon into forebrain; the remaining call types are represented in several separate areas not continuous with each other. In medulla and pons the responsive substrates for vocalization follow the course of the spinothalamic tract; in midbrain they lie within the periaqueductal gray, lateral tegmentum, and lemniscus medialis; in diencephalon they are found in the hypothalamus and midline thalamus; in forebrain, finally they are distributed over amygdala, stria terminalis, substantia innominata, preoptic region, septum, rostral hippocampus, posteromedial orbital cortex, cingulate gyrus, and rostroventral temporal cortex. Hence a close relation between limbic system and vocalization producing substrates emerges. Among the brain structures yielding vocalization the mesencephalic periaqueductal gray is assumed to be the region where the electrical stimulus interferes most directly with specific vocalization mechanisms. Besides the anatomical site of stimulation the set of stimulus parameters is important for the elicitation of vocalizations. Relations between amplitude, frequency, and duration of impulses on the one hand and type, loudness, rhythm, duration, and latency of vocalization on the other hand were tested. The influence of the stimulus set on the reaction parameters depends also on the relative position of the electrode within the effective structure. Proper manipulation of stimulus parameters often results in the disintegration of a complex stimulus response into single components.