The surface morphology and cytoarchitecture of human cingulate cortex was evaluated in the brains of 27 neurologically intact individuals. Variations in surface features included a single cingulate sulcus (CS) with or without segmentation or double parallel sulci with or without segmentation. The single CS was deeper (9.7 +/- 0.81 mm) than in cases with double parallel sulci (7.5 +/- 0.48 mm). There were dimples parallel to the CS in anterior cingulate cortex (ACC) and anastomoses between the CS and the superior CS. Flat maps of the medial cortical surface were made in a two-stage reconstruction process and used to plot areas. The ACC is agranular and has a prominent layer V. Areas 33 and 25 have poor laminar differentiation, and there are three parts of area 24: area 24a adjacent to area 33 and partially within the callosal sulcus has homogeneous layers II and III, area 24b on the gyral surface has the most prominent layer Va of any cingulate area and distinct layers IIIa-b and IIIc, and area 24c in the ventral bank of the CS has thin layers II-III and no differentiation of layer V. There are four caudal divisions of area 24. Areas 24a' and 24b' have a thinner layer Va and layer III is thicker and less dense than in areas 24a and 24b. Area 24c' is caudal to area 24c and has densely packed, large pyramids throughout layer V. Area 24c' g is caudal to area 24c' and has the largest layer Vb pyramidal neurons in cingulate cortex. Area 32 is a cingulofrontal transition cortex with large layer IIIc pyramidal neurons and a dysgranular layer IV. Area 32' is caudal to area 32 and has an indistinct layer IV, larger layer IIIc pyramids, and fewer neurons in layer Va. Posterior cingulate cortex has medial and lateral parts of area 29, a dysgranular area 30, and three divisions of area 23: area 23a has a thin layer IIIc and moderate-sized pyramids in layer Va, area 23b has large and prominent pyramids in layers IIIc and Va, and area 23c has the thinnest layers V and VI in cingulate cortex. Area 31 is the cinguloparietal transition area in the parasplenial lobules and has very large layer IIIc pyramids. Finally, variations in architecture between cases were assessed in neuron perikarya counts in area 23a. There was an age-related decrease in neuron density in layer IV (r = -0.63; ages 45-102), but not in other layers.(ABSTRACT TRUNCATED AT 400 WORDS)