Some Anatomical Bases of Cortical Somatotopic Organization; pp. 325–347

The evoked potential method was used to define the cortical visual, auditory and somatosensory areas in the mouse. Two somatotopic representations (SmI and SmII) were found which in general conform to the patterns described for other placental mammals. The representation of the face, and particularly the mys-tacial vibrissae, was greatly expanded in SmI. Nissl stained sections through the SmI face area revealed an unusual organization of the granule cells in layer IV. By correlating the appearance of layer IV as seen in sections perpendicular and parallel to the plane of layer IV, it was shown that the granule cells form cylindrical multicellular cytoarchitectonic units which, because of their shape, were called ''barrels''.The barrels were used as markers to define a cytoarchitectonic field – the barrel field – which was found to coincide with the SmI representation of the face and parts of the forepaw. Within the barrel field, a particularly consistent region was recognized – posteromedial barrel subfield (PMBSF) – in which the organization of the barrels was remarkably similar to the organization of tactile organs – the vibrissae on the contralateral face. Each PMBSF barrel is directly related to a single vibrissae since neonatal lesions of the vibrissae resulted in a predicted altered pattern in the PMBSF. And, some species which have more rows of whiskers than the mouse have more rows of PMBSF barrels.The larger PMBSF barrels contained approximately 2,000 neurons and the cytoarchitectonic basis for the barrels was shown to be purely a neuronal packing phenomenon. There was a direct proportionality between the number of myelinated fibers supplying a vibrissa and the number of neurons in the barrel to which it projects. In Golgi preparations, 85% of the layer IV neurons were found to have their dendritic processes restricted to a single barrel. From quantitative measures, Golgi impregnated barrel neurons were found to belong to only two fundamental classes: the Class I cells had smaller somata and spiny dendrites, while the Class II cells had larger somata and smooth dendrites. The specific thalamocortical projections from the ventrobasal complex to the barrels were demonstrated to terminate only within individual barrels. Silver grain counts of these projections showed that the terminal distribution within a barrel is not homogeneous.The cytoarchitectonic alterations in the PMBSF which were produced by vibrissal damage in young animals corresponded directly to an altered organization in the thalamocortical projections. There was a ''critical period'' after which it was not possible to alter PMBSF architecture, but during which there was a progressive pattern of cortical changes. Following the peripheral lesions, there was no loss of total PMBSF area, but rather barrels adjacent to the affected zone became enlarged as if to ''compensate'' for areal losses produced by the lesions.From experiments using the metabolic marker 2-deoxyglucose, it was shown that the barrels were part of a cortical columnar organization extending from pia to white matter. In addition, these experiments showed that removal of the vibrissal hairs alone was sufficient to deprive the barrel cortex of activity in behaving animals.These results are discussed in relation to some of the developmental determinants of cortical structure and possible mechanisms for the evolution of organization in established sensory pathways.