Intrinsic inter- and intralaminar connections and their relationship to the tonotopic map in cat primary auditory cortex

Small iontophoretic injections of the lectin, Phaseolus vulgaris leucoagglutinin (PHA-L), were made into different layers of the primary auditory cortex (AI) of cats. Injections in layer I labeled two types of morphologically distinct fibers in layer I as well as a smaller number of axons in layers II and III. Layer II injections labeled descending axons that produced a dense plexus of terminal fibers in layers I–III of both AI and adjacent auditory fields. Injections in layer III also labeled a dense plexus of axon collaterals at the junction of layers V and VI and labeled patches of terminal fibers in both AI and adjacent auditory fields. These were densest in layers I–III but usually extended into layers IV and V as well. The patches were partly formed by axon collaterals of layer III pyramidal cells that traveled for over 4 mm in the gray matter. Injections confined to layer IV labeled axons in all layers of the cortex but none of these axons appeared to reach the white matter. The axons spread laterally in layer IV and up into the superficial layers and ramified especially layer I. Injections in layers V and VI labeled axons in all layers of the cortex but these were densest in the deep layers where labeling was fairly homogeneous. In the upper layers the labeling was arranged in semi-discrete patches. Large injections involving layers I–III were studied in tangential sections. Between 3 and 8 patches of terminal labeling were observed in AI and these were mainly arranged in a band with its long axis aligned approximately in the dorsoventral direction. However dense patches of terminal labeling also occurred both anterior and posterior to the injection site. In selected experiments portions of the tonotopic map in AI were mapped by single unit recording and subsequently the map was related to patches of anterogradely labeled fibers that surrounded injections of PHA-L. Rows of dorsoventrally oriented patches were among cells with a similar best frequency to those in the injection site. However patches located anterior or posterior to the injection site were among cells with higher or lower best frequencies. Two injections of PHA-L close together produce different patterns of labeling. One of the injections usually produces one or more patches that has no correlate among the patches of fibers labeled by the adjacent injection. This is clearest when one of the injections is made with biotinylated PHA-L that can be visualized directly without the use of primary antibodies. Thus the intrinsic connections of AI arising from nearby cylinders of neurons are not homogenous and clusters of cells can be identified by their unique pattern of connections within AI.