Kreitewolf Jens, Mathias Samuel R, von Kriegstein Katharina
Department of Psychology, University of LübeckLübeck, Germany.
Max Planck Institute for Human Cognitive and Brain SciencesLeipzig, Germany.
Front Psychol. 2017 Sep 14;8:1584. doi: 10.3389/fpsyg.2017.01584. eCollection 2017.
Previous studies have shown that listeners are better able to understand speech when they are familiar with the talker's voice. In most of these studies, talker familiarity was ensured by explicit voice training; that is, listeners learned to identify the familiar talkers. In the real world, however, the characteristics of familiar talkers are learned incidentally, through communication. The present study investigated whether speech comprehension benefits from implicit voice training; that is, through exposure to talkers' voices without listeners explicitly trying to identify them. During four training sessions, listeners heard short sentences containing a single verb (e.g., "he writes"), spoken by one talker. The sentences were mixed with noise, and listeners identified the verb within each sentence while their speech-reception thresholds (SRT) were measured. In a final test session, listeners performed the same task, but this time they heard different sentences spoken by the familiar talker and three unfamiliar talkers. Familiar and unfamiliar talkers were counterbalanced across listeners. Half of the listeners performed a test session in which the four talkers were presented in separate blocks (blocked paradigm). For the other half, talkers varied randomly from trial to trial (interleaved paradigm). The results showed that listeners had lower SRT when the speech was produced by the familiar talker than the unfamiliar talkers. The type of talker presentation (blocked vs. interleaved) had no effect on this familiarity benefit. These findings suggest that listeners implicitly learn talker-specific information during a speech-comprehension task, and exploit this information to improve the comprehension of novel speech material from familiar talkers.
以往的研究表明，当听众熟悉说话者的声音时，他们能更好地理解言语。在大多数此类研究中，通过明确的声音训练来确保说话者的熟悉度；也就是说，听众学会识别熟悉的说话者。然而，在现实世界中，熟悉的说话者的特征是通过交流偶然习得的。本研究调查了言语理解是否受益于隐性声音训练；也就是说，通过接触说话者的声音，而听众并未明确试图识别他们。在四次训练课程中，听众听到由一名说话者说出的包含单个动词的短句（例如，“他写字”）。这些句子与噪音混合，听众在测量其言语接受阈值（SRT）的同时识别每个句子中的动词。在最后的测试环节中，听众执行相同的任务，但这次他们听到熟悉的说话者和三名不熟悉的说话者说出的不同句子。熟悉和不熟悉的说话者在听众中相互平衡。一半的听众进行了一个测试环节，其中四名说话者在单独的块中呈现（分块范式）。对于另一半听众，说话者在每次试验中随机变化（交错范式）。结果表明，当言语由熟悉的说话者发出时，听众的SRT低于不熟悉的说话者。说话者呈现的类型（分块与交错）对这种熟悉度优势没有影响。这些发现表明，听众在言语理解任务中隐性地学习说话者特定的信息，并利用这些信息来提高对来自熟悉说话者的新言语材料的理解。
