Figure 4.

A, For short-term experiments, the percentage of the learned change that was blocked was correlated with the percentage of cancellation at 0.2 Hz after bilateral injection of CNQX (circles) or PBS alone (triangles). The squares indicate where CNQX was injected for long-term experiments. Filled symbols represent data from c at H, open triangles and circles represent data from cat K, and open squares represent data from cat N. After CNQX injections in the long-term experiments, cancellation was affected but memory was not. Dashed lines emphasize the lack of overlap between conditions. B, Similar results were obtained in short-term experiments if the delay period between learning and injection was 40 min rather than 20 min.