The prefrontal cortex and oculomotor delayed response: a reconsideration of the "mnemonic scotoma"

Abstract

The concept of the "mnemonic scotoma," a spatially circumscribed region of working memory impairment produced by unilateral lesions of the PFC, is central to the view that PFC is critical for the short-term retention of information. Presented here, however, are previously unpublished data that offer an alternative, nonmnemonic interpretation of this pattern of deficit. In their study, Wajima and Sawaguchi [Wajima, K., & Sawaguchi, T. The role of GABAergic inhibiton in suppressing perseverative responses in the monkey prefrontal cortex. Neuroscience Research, 50(Suppl. 1), P3-P317, 2004] applied the GABA(A) antagonist bicuculline methiodide unilaterally to the PFC of two monkeys while they performed an oculomotor delayed-response task. Consistent with previous studies, errors for the initial memory-guided saccade were markedly higher when the cued location fell into the region of the visual field affected by the infusion. These erroneous saccades tended to select an alternative target location (out of a possible 16) that had not been cued on that trial. By extending the analysis window, however, it was observed that the second, "corrective" saccade often acquired the location that had been cued on that trial. Further analysis of the erroneous initial saccades indicated that they tended to be directed to a location that had been relevant on the previous trial. Thus, the deficit was not one of "forgetting" the cued location. Rather, it was one of selecting between currently and previously relevant locations. These findings suggest a need for a reconsideration of the concept of the mnemonic scotoma, which in turn invites a reconsideration of functional interpretations of sustained neuronal activity in PFC.

Figure 1

Delayed-response tasks, and neuropsychology and neurophysiology using such tasks. A. Schematic drawing of a classic delayed-response task administered with the Wisconsin General Test Apparatus. In the Sample phase, the monkey observes while one of the food wells is baited with a food reward. During the Delay phase, an opaque screen is lowered, and both food wells are covered by identical objects. The Response phase is initiated by the lifting of the screen, upon which the monkey selects one of the food wells by displacing the cover in order to retrieve the reward. Illustration adapted from Curtis and D’Esposito (2004). B. Sequence of a common version of the oculomotor delayed-response (ODR) task. Each rectangle shows the screen at a time during the trial. Dashed lines and arrow show the monkey’s point of fixation. C. Activity from a DLPFC neuron located in the right hemisphere. Only trials with upper left targets (135°) and those with the opposite direction (315°) are shown. From Funahashi et al. (1989) D. Reconstruction of the lesion that gives rise to the behavioral deficit shown in panel E. From Funahashi et al. (1993). E. Changes in performance on the ODR task after unilateral lesion to the DLPFC. From Funahashi et al. (1993).

Figure 2

Results from Wajima and Sawaguchi (2004), illustrating the effects of BMI injection in left DLPFC on initial-saccade accuracy in the ODR and control (CON) tasks. Data are from 60 min before injection and from 100 min after injection. A. Percentage change in the discrepancy between the position of the target and the end point of the initial saccades after drug injection, shown separately for each target location. Performance at each location is represented by the size of the boxes centered on each, with dashed lines representing 0% change, and solid lines indicating the proportion change. Shading denotes target locations for which the discrepancy increased significantly after injection (Mann-Whitney U test; p values are represented by the color code indicated to the right). Inset shows the injection site of this session, which was the left DLPFC. B: Superimposed 2-dimensional trajectories of saccades in the ODR and CON task are shown separately for pre- and post-drug periods. Trials with three target locations are shown for each line. The target locations for each line are shown in the insets (arrow). White squares show the fixation point. Abbreviations: Pre, predrug (before injection) period; Post, postdrug (after injection) period.

Figure 3

A. 2-dimensional trajectories of initial saccades in the ODR task, incorporating the same data from Fig. 2B, but with the actual (solid circle) or potential (dashed circles) target windows. Saccades falling within the solid circle correspond to correct trials. Traces with red asterisks show (erroneous) saccades that were directed toward the location that had been cued and captured in the previous trial (i.e., the previous trial was correctly performed). Traces with green and blue asterisks correspond to saccades for which the previous trial was an error trial, with green indicating saccades to the location cued on the previous trial and blue indicating saccades to the location acquired by the initial saccade on the previous trial. B. Incorporates data from A, but with the time window extended to include the saccade that followed the initial saccade. Note that most of these second saccades were directed to the location that was to be chosen by the first saccade (i.e., most of these are “corrective saccades”). C. Distribution of erroneous initial saccades on all error trials from both monkeys, illustrating quantitatively that errors showed a greater influence of information from the previous, rather than the current, trial. Each trial is scored in two ways: the difference of direction (in degrees) between the initial saccade and the location cued on the current trial (gray); and the difference of direction between the initial saccades and the location cued on the previous trial (black). Arrows indicate the bar containing the median value for each scoring procedure. (Note that this analysis does not take into account whether the previous trial was correct, and thus does not reflect the additional factor of the magnitude of the error relative to the initial saccade on the previous trial. I.e., unlike panel B, this analysis does not distinguish “proactive interference” from “perseverative” errors.) D. Percentages of erroneous saccades that were directed to the previous cue or target locations are shown separately for the target location of the previous trial (ipsi vs. contralateral visual fields to the injected site). Errors were observed more often when the previous cue had been located in the ipsilateral visual fields than when that had been presented in the contralateral visual field.