ORIGIN OF THE APPARENT TISSUE CONDUCTIVITY IN THE MOLECULAR AND GRANULAR LAYERS OF THE IN-VITRO TURTLE CEREBELLUM AND THE INTERPRETATION OFCURRENT SOURCE-DENSITY ANALYSIS

1. We determined the origin of the apparent tissue conductivity (sigma (a)) of the turtle cerebellum in vitro. 2. Application of a current wi th a known current density (J) along the longitudinal axis of a conduc tivity cell produced an electric field in the cerebellum suspended in the cell. The measured electric field (E) perpendicular to the cerebel lar surface indicated a significant inhomogeneity in sigma(a) (= J/E) with a major discontinuity between the molecular layer (0.25 +/- 0.05 S/m, mean +/- SD) and granular layers (0.15 +/- 0.03 S/m) (n = 39). 3. This inhomogeneity was more pronounced after anoxic depolarization. T he value of sigma(a) decreased to 0.11 +/- 0.03 and 0.040 +/- 0.008 S/ m in the molecular and granular layers, respectively. The ratio of sig ma(a)s in the two layers increased from 1.67 in the normoxic condition to 2.75 after anoxic depolarization. 4. This difference in sigma(a) a cross the two layers was present within the range of frequencies (DC t o 10 kHz) studied where the phase of sigma(a) was small (less than +/- 2 degrees) and therefore sigma(a) was ohmic. 5. The inhomogeneity in sigma(a) was in part due to an inhomogeneity in the extracellular cond uctivity (sigma(e)) as determined from the extracellular diffusion of ionophoresed tetramethylammonium. Like sigma(a), the value of sigma(e) was also higher in the molecular layer (0.165 S/m) than in the granul ar layer (0.097 S/m). The inhomogeneity in sigma(e) was due to a small er tortuosity and a larger extracellular volume fraction in the molecu lar layer compared with the granular layer. 6. sigma(a) was, however, consistently higher, by similar to 50%, than sigma(e). A core conducto r model of the cerebellum indicated that these discrepancies between s igma(a) and sigma(e) were attributable to additional conductivity prod uced by a passage of the longitudinal applied current through the intr acellular space of Purkinje cells and ependymal glial cells, with the glial compartment playing the dominant role. Cells with a long process and a short space constant such as the ependymal glia evidently enhan ce the effective ''extracellular'' conductivity by serving as intracel lular conduits for the applied current. The result implies that the ef fective sigma(e) may be larger than sigma(e) for neuronally generated currents in the turtle cerebellum because the space constant for Purki nje cells is several times greater than that for the ependymal glia an d consequently Purkinje cell-generated currents travel over a long dis tance relative to the space constant of glial cells. 7. Some implicati ons of the inhomogeneity were examined by comparing the depth profiles of the current source density (CSD) estimated for various conductivit y profiles. The CSD profile for the inhomogeneous case, using the meas ured profile of sigma(a), did not differ qualitatively from that for t he homogeneous case using the average of measured sigma(a)s, suggestin g that the CSD analysis may provide a qualitatively accurate picture o f actual CSD profiles even if one uses a homogeneous approximation for the actual conductivity. However, quantitatively the CSD profiles wer e different in the two cases, demonstrating importance of measurements of the conductivity profile for rigorous analyses of CSD.