Temporal frequency characteristics of synchrony-asynchrony discrimination of audio-visual signals
Temporal frequency characteristics of synchrony-asynchrony discrimination of audio-visual signals
复制标题
DOI：
10.1007/s00221-005-2385-8
复制
发表时间：
2005-10-01
影响因子：
2
通讯作者：
Nishida, S
中科院分区：
文献类型：
--
作者：
Fujisaki, W;Nishida, S
时间同步是整合不同感官方式呈现的信息的关键条件。为了深入了解视听信号同步感知的机制，我们检查了人类参与者检测同步视听刺激的时间限制。具体来说，我们在改变时间频率和/或调制波形时测量了作为视听滞后函数的同步-异步辨别的正确率百分比。视听刺激是亮度调制的高斯斑点和幅度调制的白噪声。结果表明，对于时间频率高于 4 Hz 的周期性脉冲序列，即使滞后足够大以区分单个脉冲，同步-异步区分也几乎不可能（实验 1）。这种时间限制不能归因于视觉或听觉中的外围低通滤波器（实验2），这表明时间限制反映了位于跨模式信号比较时或之前的更中心机制的属性。我们还发现该中心机制的功能行为无法用线性低通滤波器来近似（实验 3）。这些结果与以下假设一致：视听同步的感知是基于从模态内信号流中分离出的显着时间特征的比较。
Temporal synchrony is a critical condition for integrating information presented in different sensory modalities. To gain insight into the mechanism underlying synchrony perception of audio-visual signals we examined temporal limits for human participants to detect synchronous audio-visual stimuli. Specifically, we measured the percentage correctness of synchrony-asynchrony discrimination as a function of audio-visual lag while changing the temporal frequency and/or modulation waveforms. Audio-visual stimuli were a luminance-modulated Gaussian blob and amplitude-modulated white noise. The results indicated that synchrony-asynchrony discrimination became nearly impossible for periodic pulse trains at temporal frequencies higher than 4 Hz, even when the lag was large enough for discrimination with single pulses (Experiment 1). This temporal limitation cannot be ascribed to peripheral low-pass filters in either vision or audition (Experiment 2), which suggests that the temporal limit reflects a property of a more central mechanism located at or before cross-modal signal comparison. We also found that the functional behaviour of this central mechanism could not be approximated by a linear low-pass filter (Experiment 3). These results are consistent with a hypothesis that the perception of audio-visual synchrony is based on comparison of salient temporal features individuated from within-modal signal streams.