Holden Laura K, Skinner Margaret W, Holden Timothy A, Demorest Marilyn E
Department of Otolaryngology-Head & Neck Surgery, Washington University School of Medicine, St Louis, Missouri 63110, USA.
Ear Hear. 2002 Oct;23(5):463-76. doi: 10.1097/00003446-200210000-00008.
The primary objective of the study was to determine whether individual cochlear implant recipients recognize speech better with an electrical stimulation rate of 720 or 1800 pulses per second per channel (pps/ch) using the Nucleus 24 Advanced Combination Encoder (ACE) speech coding strategy. The secondary objective was to determine, for each active electrode, the relation between psychophysical measures and MAP minimum and maximum stimulation levels for each rate, as well as the stability of MAP minimum and maximum levels during the study.
Eight postlinguistically deaf adults implanted with the Nucleus 24 device participated in this study comparing the effect of a moderate (720 pps/ch) and a fast (1800 pps/ch) rate of electrical stimulation on speech recognition of words in quiet and sentences in noise presented at 50, 60, and 70 dB SPL in the laboratory and on listening to sound in everyday life over a 14-wk time period. At the beginning of the study, psychophysical measures (i.e., counted threshold and maximum acceptable loudness [MAL] levels) were obtained for each active electrode with each of the two rates to initially set MAP minimum and maximum stimulation levels. These levels were then adjusted to make speech and environmental sound clear and comfortable in everyday life. Threshold and MAL levels were obtained again half way through the study to monitor possible hearing changes. A four-phase test design for evaluation of speech recognition was followed; an equal number of subjects started with each of the two rates and alternated rates for each phase. In the last 2 wk of each phase, word and sentence scores were obtained, and subjects responded to a questionnaire. For the group, factorial analyses of variance were conducted for subject, stimulation rate, and time period (first two phases versus second two phases) for words, phonemes within words, and sentences at each level. Additional analyses were obtained for individual subjects.
Group mean scores across time periods were significantly higher for 1800 pps/ch than 720 pps/ch for phonemes and sentences in noise at 50 dB SPL. There was no significant difference in scores for phonemes and sentences at 60 and 70 dB SPL or for words at any of the three levels. Group mean scores across stimulation rate were significantly higher during the second half than the first half of the study for words, phonemes, and sentences at 50 dB SPL. This result is consistent with subjects learning to recognize speech cues near threshold. A subject by rate interaction was seen for sentences at 70 dB SPL and for all three speech measures at 50 dB SPL. These interactions reflect the fact that two subjects performed significantly better with 720 pps/ch, whereas two other subjects performed significantly better with 1800 pps/ch. Responses to the questionnaire indicated that two subjects preferred 720 pps/ch, three preferred 1800 pps/ch, and three had no preference. The minimum and/or maximum levels in most subjects' final MAPs differed from the psychophysical measures for both rates. Changes in Current Level at threshold and MAL were minimal from the first to the second half of the study for each rate.
More than half the subjects preferred one of the two rates for use in everyday life, and four subjects performed significantly better with one of the two rates on at least one test measure. These findings underscore the clinical importance of creating MAPs for each implant recipient that include at least a moderate and a fast rate within ACE during the first months of device use. Given the significant learning effects for soft speech that occurred over several weeks use of each rate in this study, it is suggested that each rate be used alone for a week or two before comparing them and deciding which provides more benefit. In addition, adjustments in an individual's MAP minimum and maximum levels are needed at each rate so soft and normal conversational speech as well as loud sound are clear and comfortable in everyday life.
本研究的主要目的是确定使用Nucleus 24高级组合编码器（ACE）语音编码策略时，人工耳蜗植入者在每秒每通道720或1800脉冲（pps/ch）的电刺激速率下是否能更好地识别语音。次要目的是确定每个有源电极在每种速率下心理物理学测量值与MAP最小和最大刺激水平之间的关系，以及研究期间MAP最小和最大水平的稳定性。
八名植入Nucleus 24设备的语后聋成年人参与了本研究，比较了中等（720 pps/ch）和快速（1800 pps/ch）电刺激速率对实验室中50、60和70 dB SPL下安静环境中的单词和噪声环境中的句子语音识别的影响，以及在14周时间内对日常生活中声音聆听的影响。在研究开始时，针对每个有源电极以两种速率中的每一种获取心理物理学测量值（即计数阈值和最大可接受响度[MAL]水平），以初步设置MAP最小和最大刺激水平。然后调整这些水平，以使日常生活中的语音和环境声音清晰舒适。在研究进行到一半时再次获取阈值和MAL水平，以监测可能的听力变化。遵循四阶段测试设计来评估语音识别；相同数量的受试者从两种速率中的每一种开始，并在每个阶段交替速率。在每个阶段的最后2周，获取单词和句子得分，受试者回答一份问卷。对于该组，对每个水平的单词、单词内音素和句子的受试者、刺激速率和时间段（前两个阶段与后两个阶段）进行方差分析。对个体受试者进行了额外的分析。
在50 dB SPL的噪声环境中，对于音素和句子，1800 pps/ch的跨时间段组平均得分显著高于720 pps/ch。在60和70 dB SPL下音素和句子的得分或在三个水平中任何一个水平下单词的得分均无显著差异。在50 dB SPL下，对于单词、音素和句子，跨刺激速率的组平均得分在研究的后半段显著高于前半段。该结果与受试者学习识别接近阈值的语音线索一致。在70 dB SPL的句子以及50 dB SPL的所有三项语音测量中观察到受试者与速率的交互作用。这些交互作用反映了这样一个事实，即两名受试者在720 pps/ch时表现明显更好，而另外两名受试者在1800 pps/ch时表现明显更好。对问卷的回答表明，两名受试者更喜欢720 pps/ch，三名受试者更喜欢1800 pps/ch，三名受试者没有偏好。大多数受试者最终MAP中的最小和/或最大水平与两种速率下的心理物理学测量值不同。对于每种速率，从研究的前半段到后半段，阈值和MAL处的电流水平变化最小。
超过一半的受试者在日常生活中更喜欢两种速率中的一种，并且四名受试者在至少一项测试测量中在两种速率中的一种下表现明显更好。这些发现强调了在设备使用的头几个月为每个植入接受者创建MAP的临床重要性，其中至少包括ACE中的中等和快速速率。鉴于在本研究中每种速率使用几周后出现了对轻声语音的显著学习效应，建议在比较并决定哪种速率更有益之前，每种速率单独使用一两周。此外，每种速率都需要调整个体的MAP最小和最大水平，以便在日常生活中轻声和正常对话语音以及大声声音都清晰舒适。
