Interactions between gaze-evoked blinks and gaze shifts in monkeys

Abstract

Rapid eyelid closure, or a blink, often accompanies head-restrained and head-unrestrained gaze shifts. This study examines the interactions between such gaze-evoked blinks and gaze shifts in monkeys. Blink probability increases with gaze amplitude and at a faster rate for head-unrestrained movements. Across animals, blink likelihood is inversely correlated with the average gaze velocity of large-amplitude control movements. Gaze-evoked blinks induce robust perturbations in eye velocity. Peak and average velocities are reduced, duration is increased, but accuracy is preserved. The temporal features of the perturbation depend on factors such as the time of blink relative to gaze onset, inherent velocity kinematics of control movements, and perhaps initial eye-in-head position. Although variable across animals, the initial effect is a reduction in eye velocity, followed by a reacceleration that yields two or more peaks in its waveform. Interestingly, head velocity is not attenuated; instead, it peaks slightly later and with a larger magnitude. Gaze latency is slightly reduced on trials with gaze-evoked blinks, although the effect was more variable during head-unrestrained movements; no reduction in head latency is observed. Preliminary data also demonstrate a similar perturbation of gaze-evoked blinks during vertical saccades. The results are compared with previously reported effects of reflexive blinks (evoked by air-puff delivered to one eye or supraorbital nerve stimulation) and discussed in terms of effects of blinks on saccadic suppression, neural correlates of the altered eye velocity signals, and implications on the hypothesis that the attenuation in eye velocity is produced by a head movement command.

Fig. 1

Likelihood of observing gaze-evoked blinks with gaze shifts. (a) The probability of generating a gaze-evoked blink is plotted as a function of gaze amplitude during head-restrained saccades (open symbols, dashed traces) and coordinated eye-head movements (filled symbols, solid traces). Data were grouped into 10° bins according to the absolute value of gaze amplitude. The probability of trials with gaze-evoked blinks is indicated on the ordinate axis. (b) Legend identifies each symbol, color, and line-type with an animal and head mobility condition. This convention is used in several figures that follow.

Fig. 2

Relationship between main sequence property and blink likelihood. (a,b) Average gaze velocity is plotted as function of amplitude for head-restrained saccades (left) and head-unrestrained gaze shifts (right). Data were grouped into 10° bins according to gaze amplitude. Each point denotes the mean with one standard deviation errorbars of the average velocity of trials without gaze-evoked blinks. (c,d) Blink probability is plotted as a function of average gaze velocity for large amplitude gaze shifts, i.e., for the range for which gaze velocity remains relatively constant for each animal. See Fig. 1b for assignment of symbols.

Fig. 3

Effects of gaze-evoked blinks on head-restrained saccades. Each panel illustrates individual trials of eye velocity (deg/s) and eyelid position (a.u.: arbitrary units) traces as a function of time. Saccades produced with (blue) and without (red) accompanying gaze-evoked blinks are differentiated by color. The mean amplitude of saccades and the number of trials shown in each panel are indicated in the inset text. Note, however, that the ratio of displayed red and blue trials need not correspond to the probability of observing a gaze-evoked blink for the movement amplitude. All plots are aligned on movement onset. The data included in the left column correspond to a smaller amplitude movement than the trials plotted in the right column. The dashed, horizontal line in gray marks zero deg/s. Data from the different animals are shown in the different rows: (a) monkey WL, (b) monkey WY, (c) monkey BN, (d) monkey BL.

Fig. 4

Effects of gaze-evoked blinks on head-unrestrained gaze shifts. Data are in the same format as Fig. 3, except that head velocity profiles are also included in each panel.

Fig. 5

Main sequence analysis for movements with (blue) and without (red) gaze-evoked blinks. Left column plots peak gaze velocity as a function of saccade amplitude for all head-restrained trials. Note that for head-restrained data, peak eye velocity equals peak gaze velocity. Middle column shows peak gaze velocity as a function of gaze amplitude for head-unrestrained data. Right column displays peak velocity of the eye component as a function of the saccadic eye component of head-unrestrained gaze shifts. Each row shows data from one animal, as identified by the two letters in the bottom right corner of the panels in the right column. Each point denotes one trial and data from all horizontal movements (see Table 1) are illustrated. The ‘x’ denotes the bin, tested in 10° increments, for which the main sequence parameter was significantly different between blink and non-blink trials (two-tailed t-test, P<0.05).

Fig. 6

A comparison of average velocity as a function of amplitude parameters for movements with (blue) and without (red) gaze-evoked blinks. Average velocity was computed as amplitude divided by duration. The illustration is in the same format at Fig. 5.

Fig. 7

A comparison of movement duration as a function of amplitude parameters for gaze shifts with (blue) and without (red) gaze-evoked blinks. The illustration is in the same format at Fig. 5.

Fig. 8

Effects of gaze-evoked blinks on large amplitude coordinated eye-head movements grouped according to head amplitude. All head-unrestrained gaze shifts with saccadic eye amplitude greater than 32° and in the same direction (leftward for monkey WL (a,b) and rightward for monkey BL (c,d)) were included in the analysis. Trials that met the inclusion criterion were grouped according to head amplitude, as identified by the color and text in insets, and sorted into trials with and without gaze-evoked blinks. The panels show temporal profiles of averaged eye-in-head and head-in-space velocities for these head-unrestrained gaze shifts without (a,c) and with (b,d) gaze-evoked blinks. Comparable plots for monkeys BN and WY are not illustrated because there were essentially no head-unrestrained trials without gaze-evoked blinks that met the inclusion criteria.

Fig. 9

Distribution of blink times. (a) For each head-restrained saccade accompanied by a gaze-evoked blink for monkey WL, the time of blink onset relative to gaze onset is plotted as a function of saccade amplitude. Negative latency value means that blink onset lags gaze onset, and positive metric indicates that blink onset leads gaze onset. The color axis in panel b defines the initial gaze position for each trial, which also equal initial eye-in-head position for head-restrained data. (b) Same display format for head-unrestrained data obtained in monkey WL. The color of each dot indicates initial gaze position, since the eyes were initially centered in the orbits prior to the onset of the gaze shift. (c,d) For each animal, the data were sorted into 5° bins of gaze amplitude. The mean±one standard deviation of blink time relative to gaze onset is plotted as a function of gaze amplitude for head-restrained (c) and head-unrestrained (d) movements. The color symbols follow the convention established in Fig. 1b.

Fig. 10

Effect of blink timing on velocity waveform. Eye position (top), eye velocity (middle), and eyelid velocity (bottom) are plotted as a function of time. The gray trace is an average of several, same amplitude head-restrained saccades produced without a gaze-evoked blink. Note that the eye velocity is a bell-shaped profile with a slightly prolonged deceleration phase, and the eyelid velocity trace remains relatively stable across the duration of the saccade. Individual, amplitude-matched, gaze-evoked saccade trials are shown in color. Corresponding traces of eye and eyelid movements are shown in the same color. The effect of a gaze-evoked blink is to induce a gross perturbation in the position and velocity traces. Furthermore, the timing of the blink contributes to the time at which the eye trajectory is compromised and also to the magnitude of attenuation in eye velocity.

Fig. 11

Effects of gaze-evoked blinks on vertical head-restrained saccades. Eye position and velocity and eyelid position and velocity (top to bottom) waveforms are plotted as a function of time for individual amplitude-matched saccades. Saccades without blinks are shown in red and movements with gaze-evoked blinks are plotted in blue.

Fig. 12

Effects of gaze-evoked blinks on latency. The data were sorted according to target configuration, such that the initial and final target positions were identical for all trials in each dataset. (a) The average gaze latency for movements with gaze-evoked blinks is plotted against movements without accompanying blinks (nonblink condition) for head-restrained saccades. Each point represents the mean value obtained from each dataset. (b) The difference in latency values (blink-nonblink conditions) is plotted as a function of desired gaze amplitude. (c,d) Same format at (a,b) but for gaze latency of head-unrestrained movements. (e,f) Same format as (a,b) but for latency of head component of head-unrestrained gaze shifts. Comparable analysis for the eye component of gaze shifts is not provided as the results were nearly identical to the gaze latency data (c,d). The color of each symbol identifies the animal, following the notation introduced in Fig. 1b. The open and filled symbol convention, however, was not applied here. Filled symbols instead indicate statistically significant difference in latency between the blink and nonblink dataset (two-tailed t-test, P<0.05). Diagonal dashed line denotes unity slope, and horizontal dashed line marks zero difference in latency.

Fig. 13

Effects of gaze-evoked blinks on movement accuracy. (a) The difference in the average horizontal amplitude (blink – nonblink trials) is plotted as a function of desired gaze amplitude for head-restrained saccades. Positive value indicates larger amplitude during gaze shifts accompanied by gaze-evoked blinks. Each point represents the mean value obtained from each dataset parsed according to target configuration (see Fig. 12). Similar analyses performed on head-unrestrained data are shown for average difference in gaze amplitude (b), ocular saccade component (c) and head amplitude (d). The color and open/filled symbol convention is the same as that used in Fig. 12. The horizontal dashed lines mark zero difference in horizontal component amplitude.

Fig. 14

Effects of gaze-evoked blinks on head movement kinematics. (a) The difference in the average horizontal peak head velocity (blink – nonblink trials) is plotted as a function of desired gaze amplitude. Positive value indicates bigger peak velocity during gaze shifts accompanied by gaze-evoked blinks. Each point represents the mean value obtained from each dataset parsed according to target configuration (see Fig. 12). (b) Same analysis for the time of peak head velocity. Positive value indicates that the head velocity peaks later when the gaze shift is accompanied by a gaze-evoked blink. The color and open/filled symbol convention is the same as that used in Fig. 12. The horizontal dashed lines mark zero difference in head kinematic.