HIPPOCAMPAL UNIT ACTIVITY IN A VISUAL SHORT-TERM MEMORY TASK IN RHESUS MACAQUES

W.E. Skaggs^1,2*; B.L. McNaughton¹; M. Permenter²;
J. Vogt²; M. Archibeque^1,2; C.A. Barnes¹

1. NSMA, Univ Arizona, Tucson, AZ, USA
2. CNPRC, Davis, CA, USA

Recordings were made from the hippocampus of two female rhesus macaques, 31 and 15 yrs old, using a chronically implanted 12-tetrode array, with accurate histological identification of recording locations. The monkeys performed a Passive Paired Associate Task (Erickson and Desimone, 1999), obtaining reward for bar-releasing to 5 out of a set of 10 presented images (1000-1500 trials/session). Five seconds prior to the test image there appeared a "paired" image, which predicted the following test image. Every few days, a new set of picture-pairs was introduced. In both animals, most of the units isolated in the CA1 and CA3 layers (old monkey, n=744; young monkey, n=370) fired at rates below 1 Hz, in a bursty manner (putative pyramidal cells). Consistent with other reports (e.g., Richies et al., 1991; Rolls et al., 1993; Colombo and Gross, 1994), over 90% of pyramidal cells were unresponsive to the visual stimuli in this task. Moreover, responsive cells exhibited little or no stimulus selectivity. Two main patterns of visual responsiveness were observed. Class 1 cells responded to both sample and choice stimuli without change in firing rate or latency across trials. Class 2 cells showed a systematic, intra-session response shift. For the first several hundred trials, they responded like Class 1 cells, after which response rate decreased, latency increased, and the cells started to fire preferentially to the "go" stimuli. Class 1 cells were observed in both monkeys, but Class 2 was only observed in the old monkey. There was no correlation between the response shift in Class 2 cells and performance. Class 1 responses were similar to those described for rat hippocampal cells in an odor-based DNMS task (Otto & Eichenbaum, 1992). These data suggest that control over primate hippocampal cell firing by visual and task-related variables can be modulated in complex, but systematic ways.

Support Contributed By: AG03376