Fig. 3 | Nature Communications

Fig. 3

From: Memory-guided microsaccades

Fig. 3

Memory-guided microsaccades had better directional accuracy than either microsaccades during fixation or corrective, visually-guided microsaccades. a We plotted the angular distribution of microsaccade direction error (that is, the difference between target direction and microsaccade direction; Methods). In this measure, if a microsaccade has the same direction as the target direction, then the movement has zero direction error. In this panel, this measure is shown for all microsaccades from monkey N occurring in the final 250 ms of fixation before fixation spot disappearance. Microsaccades could occur either towards or away from the remembered target location. The histogram was normalized by the total number of observations. b The same monkey exhibited highly accurate microsaccade directions toward the remembered target location when explicitly instructed to do so. c After target reappearance, corrective, visually-guided microsaccades were also target-directed, but their direction errors were significantly more variable than the instructed movements in b (see statistics in text). Thus, instructed memory-guided microsaccades were highly directionally accurate. d Each group of colored arrows shows the directions of 40 randomly selected memory-guided microsaccades when remembered target directions were drawn from the underlying corresponding-colored faint cone. Directionally accurate microsaccades could be generated for all oblique target directions. eh Similar observations for monkey M. il Similar observations for monkey P. mp Similar observations for our human subjects. In all cases, directional accuracy was the highest for the instructed, memory-guided movements and not for fixational microsaccades (a, e, i, m) or corrective, visually guided ones (c, g, k, o); also see Supplementary Fig. 3 for flash-induced microsaccades. For the leftmost column, n = 1,432, 2,875, 4,870, 2,831 microsaccades for monkey N, monkey M, monkey P, and the human subjects, respectively. The numbers of trials for memory-guided or corrective, visually guided microsaccades are the same as in Fig. 2

Back to article page