Woldorff M G, Hillyard S A
Department of Neurosciences, University of California, San Diego, La Jolla 92093-0608.
Electroencephalogr Clin Neurophysiol. 1991 Sep;79(3):170-91. doi: 10.1016/0013-4694(91)90136-r.
Two dichotic listening experiments were performed in which stimulus and task conditions were optimized for the early selection of inputs. Subjects listened selectively to sequences of rapidly presented tone pips in one ear while ignoring tone pips of a different pitch in the opposite ear. In both experiments, an enhanced positivity between 20 and 50 msec (the 'P20-50') was observed over central and frontal sites in the ERPs to the attended-channel tone pips. At longer latencies, the effects of attention appeared to include an amplitude modulation of several exogenous ERPs, including subcomponents of the central N1 (60-150 msec) and P2 (170-230 msec) waves and the temporal T complex (80-150 msec). In contrast, the attention effect prefrontally consisted of a broad negativity that appeared to be largely endogenous. A signal processing technique (Adjar) was employed to correct for distortion of mutually overlapping ERPs elicited by successive stimuli presented at short interstimulus intervals (ISIs). It was confirmed that the P20-50 attention effect was not the result of differential overlap from previous ERPs. In addition, this technique allowed an analysis to be made of the effects of the preceding stimulus type and ISI on the attention-sensitive ERPs, which provided further support for the view that highly focused selective attention can directly modulate exogenous components of the auditory ERP. Moreover, these sequence-dependent ERP modulations were paralleled by variations in target discrimination performance. Taken together, these results provide strong support for the early selection hypothesis that attention can serve to selectively bias or gate stimulus processing before full perceptual analysis has occurred.
进行了两项双耳分听实验，其中刺激和任务条件针对输入的早期选择进行了优化。受试者选择性地聆听一只耳朵中快速呈现的音调序列，同时忽略另一只耳朵中不同音高的音调。在两项实验中，在事件相关电位（ERP）中，针对被关注通道的音调，在中央和额叶部位观察到20至50毫秒之间增强的正电位（“P20 - 50”）。在更长的潜伏期，注意力的影响似乎包括对几种外源性ERP的幅度调制，包括中央N1波（60 - 150毫秒）和P2波（170 - 230毫秒）的子成分以及颞叶T复合波（80 - 150毫秒）。相比之下，前额叶的注意力效应由一个广泛的负电位组成，这似乎在很大程度上是内源性的。采用了一种信号处理技术（Adjar）来校正由短刺激间隔（ISI）呈现的连续刺激引发的相互重叠ERP的失真。证实了P20 - 50注意力效应不是先前ERP差异重叠的结果。此外，该技术允许分析先前刺激类型和ISI对注意力敏感ERP的影响，这为高度集中的选择性注意力可以直接调节听觉ERP的外源性成分这一观点提供了进一步支持。而且，这些与序列相关的ERP调制与目标辨别性能的变化并行。综上所述，这些结果为早期选择假说提供了有力支持，即注意力可以在完全感知分析发生之前选择性地偏向或控制刺激处理。
