Bair W, O'Keefe L P
Howard Hughes Medical Institute, Center for Neural Science, New York University, New York 10003, USA.
Vis Neurosci. 1998 Jul-Aug;15(4):779-86. doi: 10.1017/s0952523898154160.
We analyzed the relationship between eye movements and neuronal responses recorded from area MT in alert monkeys trained to maintain visual fixation during the presentation of moving patterns. The monkeys made small saccades which moved the eyes with velocities that spanned the sensitivity range of MT neurons. The saccades evoked changes in the neuronal response that depended upon (1) the level of stimulus-evoked activity amidst which the saccade occurred and (2) the direction of the saccade relative to the preferred direction of the neuron. Most notably, saccades were able to suppress stimulus-evoked activity when they caused retinal image flow that opposed the neuron's preference and were able to elicit a response or enhance weak activity when they caused flow in the neuron's preferred direction. On average, the disturbance lasted 40 ms beginning about 40 ms following saccade onset. Using these parameters, we simulated synthetic spike trains from an imaginary pair of similarly tuned neurons and determined that the interneuronal correlation due to saccades should be negligible at all but the lowest ongoing firing rates. This conclusion was supported from our data by the observation that response variance for single MT spike trains was not measurably reduced during periods of stable gaze compared to periods when eye movement exceeded a stability criterion (0.1 deg during 0.5 s). While the intrusions caused by saccades are too short-lived and infrequent to account for the variability of MT neuronal response (counter to the finding in V1 of Gur et al., 1997), the clear directional signal that they carry in area MT suggests that motion perception is not blocked during saccades by suppression at early stages in the visual pathway.
我们分析了警觉猴子在呈现运动模式时保持视觉注视训练过程中,眼动与从MT区记录的神经元反应之间的关系。猴子进行小幅度扫视,其眼睛移动速度跨越了MT神经元的敏感范围。扫视引发的神经元反应变化取决于:(1)扫视发生时刺激诱发活动的水平;(2)扫视方向相对于神经元偏好方向的关系。最值得注意的是,当扫视引起的视网膜图像流动与神经元偏好相反时,扫视能够抑制刺激诱发的活动;而当扫视引起的流动方向与神经元偏好方向一致时,扫视能够引发反应或增强微弱活动。平均而言,干扰在扫视开始后约40毫秒开始,持续40毫秒。利用这些参数,我们从一对假想的类似调谐神经元模拟了合成脉冲序列,并确定除了最低的持续放电率外,扫视引起的神经元间相关性应可忽略不计。我们的数据支持了这一结论,观察发现,与眼动超过稳定性标准(0.5秒内0.1度)的时期相比,在稳定注视期间单个MT脉冲序列的反应方差没有明显降低。虽然扫视引起的干扰持续时间太短且不频繁,无法解释MT神经元反应的变异性(与Gur等人1997年在V1区的发现相反),但它们在MT区携带的清晰方向信号表明,在扫视期间,视觉通路早期的抑制作用不会阻断运动感知。