1. Two monkeys were trained to fixate at a point on the monitor and to maintain fixation for 0 to 2000 ms. After cue time which was the offset of the fixation point the monkey had to saccade toward the figure location. The maximum time allowed for responding to the figure location was 500 ms. Trials where eye position left the fixation window (1° by 1°) before cue time were discarded. Eye movements were monitored by using scleral search coils (5). Stimuli were presented on a 21-inch monitor screen (28° by 21° of visual angle resolution 1024 by 768 pixels refresh rate 72.34 Hz). In each trial a red fixation spot (0.2°) popped up in a prestimulus texture consisting of random pixels with a 50% probability of being either black or white. After the monkey had fixated this spot for 300 ms the random dots were moved very briefly (28 ms i.e. two video frames) and over a very short distance (two pixels 0.06°). Motion onset was considered stimulus onset. The direction of movement of the dots was 45° 135° 225° or 315°. The movement of a square “figure” region (3° of visual angle) was always 180° opposite to that of the background. Motion-defined figures could (randomly) appear at one of three possible locations (eccentricity. 2.7° to 4.4°)
2. Multiunit activity was recorded through 16 microwire electrodes in each monkey (4 6). Receptive-field size ranged from 0.7° to 1.4° (median 0.9°) and eccentricity from 1.3° to 2.8° in one monkey and from 3.4° to 5.7° in the other. For each monkey figure positions and electrodes were chosen such that the figures covered the receptive fields of 16 electrodes simultaneously.
3. Data were obtained in pseudo-randomly interleaved blocks of trials in 9 to 12 daily sessions. To ensure identical receptive-field stimulation for figure and background responses we compared the responses to the same directions of motion for figure and ground and averaged the responses to the different directions of motion. Electrophysiological data obtained during the interval between stimulus onset and cue time were analyzed. Response strength was calculated as the average normalized activity during the intervals. Contextual modulation was calculated by subtracting the background response strength from the figure response strength. For all statistics paired t tests were applied.
4. Figure-ground activity in primary visual cortex is suppressed by anesthesia
5. V. A. F. Lamme H. Supèr R. Landman P. R. Roelfsema H. Spekreijse Vision Res. 40 1507 (2000).