Sequence regularity and geometry of orientation columns in the monkey striate cortex.

The striate cortex of the macaque monkey is subdivided into two independent and overlapping systems of columns termed “orientation columns” and “ocular dominance columns.” The present paper is concerned with the orientation columns, particularly their geometry and the relationship between successive columns. The arrangement of the columns is highly ordered; in the great majority of oblique or tangential microelectrode penetrations the preferred orientations of cells changed systematically with electrode position, in a clockwise or counterclockwise direction. Graphs of orientation vs. electrode track distance were virtually straight lines over distances of up to several millimeters; such orderly sequences were often terminated by sudden changes in the direction of orientation shifts, from clockwise to counterclockwise or back. The orientations at which these reversals occurred were quite unpredictable. Total rotations of 18% 360” were frequently seen between reversals. In tangential or almost tangential penetrations orientation shifts occurred almost every time the electrode was moved forward, indicating that the columns were either not discrete or had a thickness of less than 25-50 p, the smallest order of distance that our methods could resolve. In penetrations that were almost perpendicular to the surface, the graphs of orientation vs. track distance were relatively flatter, as expected if the surfaces of constant orientation are perpendicular to the cortical surface. S tep wise changes in orientation of about 10” could sometimes be seen in perpendicular penetrations, each orientation persisting through several clear advances of the electrode, suggesting a set of discrete columnar subdivisions. The possibility of some kind of continuous variation in orientation with horizontal distance along the cortex was not, however, completely ruled out. Occasionaliy a highly ordered sequence was broken by an abrupt large shift in orientation of up to 90”. Shifts in ocular dominance occurred roughly every 0.25-0.5 mm and were independent of orientation shifts. In multiple parallel penetrations spaced closer than about 250 p the slopes of the orientation vs. track distance curves were almost the same; reconstruction of these penetrations indicated that the regions of constant orientation are parallel sheets. On crossing perpendicular to these sheets, a total orientation shift of 180” took place over a distance of 0.51.0 mm. Column thickness, size of shifts in orientation, and the rate of change of orientation with distance along the cortex seemed to be independent of eccentricity, at least between 2 ” and 15” from the fovea. A few penetrations made in area 17 of the cat and in area 18 of the monkey showed similar orderly sequences of receptive-field orienta-