Enhanced performance with brain stimulation: attentional shift or visual cue?

Abstract

The premotor theory of visual spatial attention proposes that the same brain activity that prepares for saccades to one part of the visual field also facilitates visual processing at that same region of the visual field. Strong support comes from improvements in performance by electrical stimulation of presaccadic areas, including the frontal eye field and superior colliculus (SC). Interpretations of these stimulation experiments are hampered by the possibility that stimulation might be producing an internal visual flash or phosphene that attracts attention as a real flash would. We tested this phosphene hypothesis in the SC by comparing the effect of interchanging real visual stimuli and electrical stimulation. We first presented a veridical visual cue at the time SC stimulation improved performance; if a phosphene improved performance at this time, a real cue should do so in the same manner, but it did not. We then changed the time of SC visual-motor stimulation to when we ordinarily presented the veridical visual cue, and failed to improve performance. Last, we shifted the site of SC stimulation from the visual-motor neurons of the SC intermediate layers to the visual neurons of the superficial layers to determine whether stimulating visual neurons produced a larger improvement in performance, but it did not. Our experiments provide evidence that a phosphene is not responsible for the shift of attention that follows SC stimulation. This added evidence of a direct shift of attention is consistent with a key role of the SC in the premotor theory of attention.

Figure 1.

Change detection tasks. A, Shifting attention with a veridical visual cue. During initial fixation, a veridical visual cue on 50% of trials indicated where the target would appear. The cue disappeared, and three patches of random dot motion appeared: one target and two distractors. Motion in the patches commenced in two stages that were either contiguous (change-visible) or separated by a 150 ms blank (change-blind). On 65% of trials, the dots in the target patch changed direction between the first and second stages of motion. The monkey's task was to determine whether the target changed direction. If the motion in the target changed direction, the correct response was a saccade to the target location. If the target did not change, the correct response was to remain fixating. The direction of motion in the distractor patches never changed. B, Shifting attention with collicular stimulation. The sequence of stages in this task was nearly identical save for the absence of the veridical visual cue. Instead of the cue, the intermediate layers of the SC were stimulated just at the time when the monkey was required to attend to the target to determine whether it changed. Stimulation occurred on 50% of the trials when the target location overlapped the SC stimulation site. C, Tests of the phosphene hypothesis. The matrix illustrates interchanging of the veridical visual cue with collicular stimulation. Each diagram schematizes the base experiment or one of the four test experiments. Within a column (for the base experiment and tests 1–3), the timing of the stimulation (left column) or veridical visual cue (right column) is interchanged. Along a row, the timing remains the same, but the stimulation and veridical visual cue are interchanged at either the time of the change (top row) or in the premotion period (bottom row). The final test, stimulation of the SC superficial layers, appears separate from the matrix.

Figure 2.

Base effects of SC stimulation on performance. A, In the base SC stimulation paradigm, we stimulated the superior colliculus around the time when the direction of dot motion in the target might change. Stimulation began 150 ms before the first stage of dot motion ended and lasted a total of 600 ms. This was true for both the change-blind and change-visible tasks. B, Sample result from a single collicular location. The proportions of hits are plotted against false positives. The gray symbols indicate the monkey's performance without collicular stimulation, and the black symbols show performance with stimulation. The connected pair of open symbols shows performance on change-blind trials, and the pair of filled symbols shows the result from change-visible trials. C, Difference (stim − no stim) in hits and false positives (Δh and Δfp, respectively) for each of eight collicular locations in two monkeys. Open symbols are from change-blind trials, and filled symbols are from change-visible trials. Positive values indicate that hits (or false positives) increased with stimulation. Results from the example SC location featured in B are indicated by squares. D, Z-scores for differences in performance shown in C. The shaded bands indicate regions where the data fail to achieve significance (p > 0.01). Points lying outside the horizontal gray area denote significant changes in hits, and points outside the vertical hatched area show significant changes in false positives. Again, the sample results from B are indicated by squares.

Figure 3.

Test 1: veridical visual cue presented during the change. A, The sequence of steps is similar to those in the base stimulation experiment (Fig. 2A), except that a veridical visual cue (for only one motion patch location) replaces the collicular stimulation on 50% of trials on which the target overlapped the visual field position for that SC location. The veridical visual cue came on at the same time as had the SC stimulation; we made no attempt to allow for the 40–50 ms time for a veridical visual cue signal to arrive at the SC, because the cue still appeared early enough (150 ms before the change). B, Change in the proportion of hits versus the change in the proportion of false positives when the veridical visual cue replaces SC stimulation. Open symbols represent change-blind trials, and filled symbols represent change-visible trials. Base data from SC stimulation (from Fig. 2C) are in the background in gray. C, Differences in proportions plotted as the z-scores. Note that many individual changes lie outside the shaded regions and are therefore significant changes (p < 0.01).

Figure 4.

Test 2: SC stimulation in the premotion period. Same schema as in Figure 3. A, Sequence of steps showing SC stimulation of the intermediate layers in the premotion period (when the veridical visual cue normally appeared) rather than at the time of motion change. B, Change in the proportion of hits versus the change in the proportion of false positives from stimulation. Open symbols represent change-blind trials, and filled symbols represent change-visible trials. The shaded symbols in the background are the data from the base experiment in Figure 2C. Note that receptive field location and stimulation threshold criteria were only achieved at six of the eight SC locations (see Materials and Methods). C, The differences in proportions from B plotted as z-scores of the changes. The only significant (p ≤ 0.01) changes were two slightly significant reductions in false positives (points to the left of the hatched region).

Figure 5.

Test 3: cueing a single target location in the premotion period. Same schema as in Figures 3 and 4. A, The sequence of steps shows the veridical visual cue (for a single target location) in the premotion period. The cue appeared on 50% of trials when the target overlapped the visual field position for the current SC location. B, Change in the proportion of hits versus the change in the proportion of false positives caused by the veridical visual cue. Open symbols represent change-blind trials, and filled symbols represent change-visible trials. Comparison data from the base experiment appear in gray in the background. C, Changes plotted as z-scores. Note that many changes in hits and false positives lie outside one or both shaded regions and are therefore significant changes.

Figure 6.

Test 4: stimulation of visual neurons in SC superficial layers. A, Sequence of steps in the SC stimulation paradigm was identical to that used in the base experiment in the SC intermediate layers (Fig. 2A). B, Changes in hits and false positives (stim − no stim) for each of six SC locations in the superficial layers. Open symbols represent change-blind trials, and filled symbols represent change-visible trials. The base data from intermediate layer stimulation are shown in gray in the background for comparison. C, z-scores for differences in performance shown in B. Note the number of points falling in the shaded zones of no significant difference (p > 0.01).

Figure 7.

Summary of results in the base experiment and four tests of the phosphene hypothesis. Using the same scheme as in Figure 1C, within a column, the time of the stimulation (left) or veridical visual cue (right) were interchanged. Along a row, the timing remains the same, but the stimulation and veridical visual cues were interchanged at either the time of the change (top row) or at the premotion time (second row). Test 4 (superficial layer stimulation) appears separate from the matrix. For each of the base and four tests, we plotted the mean and SE for the change in hits (Δh) versus the mean and SE of the change in false positives (Δfp). Each point shows the mean for one experiment pooled over all SC locations. Gray dots and lines show means and SEs for each test and for comparison are repeated on all of the graphs. For a given experiment, the relevant mean and SE are plotted in black in the appropriate box and are further indicated by a black line from the origin to the mean for that experiment. The scale for the axes appears in the bottom right. None of our tests produced the same type of change in performance as did stimulation of the SC intermediate layers. See summary in Results for details.