In a recent letter to the editor (Liu and Bilkey, 1998), we addressed the question of whether the perirhinal cortex (PRC) projects directly to the hippocampus. On the basis of the findings of Liu and Bilkey (1996, 1997) and a number of previous anatomical studies (cited in this previous letter), we concluded: (1) that the evidence for a direct projection from the perirhinal cortex to the DG was somewhat equivocal and (2) that there was evidence to support our proposition that there was a direct projection from the perirhinal cortex to the CA1 region of the dorsal hippocampus. We then suggested, however, that the issue was not as yet fully resolved and that more extensive work would be required in which the border between entorhinal cortex and perirhinal cortex was clearly defined and the topographical relationship between the perirhinal cortex and hippocampus was explored in a single study.
We are pleased to see that this work is in progress, as described in the letters of both Witter et al. (1999) and Canning and Leung (1999). These authors describe their own evidence, some as yet unpublished, and suggest that we may have reached an erroneous conclusion. Their main argument is that we have misinterpreted our CSD findings, and that our ''perirhinal'' stimulation may actually have been activating the adjacent portion of the lateral entorhinal cortex (LEC) or the ventral lateral perforant path (LPP). Although we can do little to assess this claim without seeing details of their data, we can address several of the issues that were raised. First, Canning and Leung suggest that our localization of the ''CA1'' sink may have been misplaced due to several factors, including electrode bending. We believe that this is unlikely, on the basis that we had a clearly defined reference to regions immediately above and below the fissure via audio monitoring of the CA1 and DG cell layers of the hippocampus. Canning and Leung also suggest that our use of ''D4'' rather than ''D2'' smoothing in our CSD may have contributed to a ''spillover'' of the molecular layer sink into the distal dendrites of CA1. Although it is true that the D4 formula will alter the apparent dorsal/ventral extent of the source, it will not alter the position of the peak value (confirmed in D2/D4 comparisons conducted on our own data). Our data consistently showed that this peak occurred at the fissure, as one would expect if there were current sinks located both above and below this region. Canning and Leung also suggest that we should expect to see a current source dorsal to our putative CA1 sink if it is of CA1, rather than DG, origin. The intensity of a more dorsal source will, however, depend on such factors as stimulation intensity, the patterns of current flow along the CA1 dendrites, and, as they suggest, smoothing techniques. We recognized this fact and were, therefore, not unduly perturbed when we failed to locate this source in our CSDs, particularly as we often failed to observe a source at this location following stimulation of the LPP projection to the hippocampus.
In order to accept the conclusion that the perirhinal cortex has minimal projections to the hippocampus, one needs, however, to be prepared to take account of all of the data. In addition to addressing the CSD findings, one