Ultrastructural analysis of somatostatin‐immunoreactive neurons and synapses in the temporal and occipital cortex of the macaque monkey

Somatostatin‐containing neurons and terminals have been analyzed in monkey temporal and occipital cortex by using light and electron microscopic immunohistochemistry. An antibody against Somatostatin‐28(1–12), that was shown previously preferentially to label fibers (Morrison et al.: Brain Research 262:344–351, 1983), was utilized. As expected, few cell bodies were labeled. At the electron microscopic level, labeled cells presented a characteristic asymmetric position of the nucleus and very few symmetric or asymmetric synapses on the somatic surface. In all areas examined, somatostatin fibers formed a dense plexus in the most superficial layers (I‐upper III). The density of labeled fibers in intermediate (deep III–IV) and deep layers (V–VI) varied considerably among areas. The synaptic relationships of the immunoreactive fibers were analyzed and postsynaptic targets quantified in V1, V2, and the superior and inferior temporal gyrus (STG and ITG, respectively). The synapses formed by somatostatin‐labeled boutons were of the symmetric type (type II) and the primary postsynaptic targets were dendritic shafts. No regional differences were found in the distribution of the postsynaptic targets in layers I‐upper III. The pattern of synapses in the deep layers was examined in STG. The frequency and distribution of postsynaptic targets was similar to the superficial layers of STG and the other temporal and occipital regions. In intermediate layers of the temporal cortex areas there was an increase in the proportion of synapses on dendritic spines. In a correlated light and electron microscopic analysis we examined synapses made by radial fibers in these regions and found that although the main targets are distal dendritic shafts, almost 40% of synapses were on dendritic spines. We suggest that the radial fibers may originate from a specialized interneuron, previously described as the double bouquet cell, and that this particular subset of somatostatin‐containing double bouquet cells is likely to exhibit a very high degree of regional heterogeneity with a preference for association cortices with extensive cortico‐cortical convergence.