Twenty listeners with normal audiometric thresholds (age, 18–29 years old; 6 males and 14 females) participated in the experiment. None of the listeners had prior experience of participating in psychophysical experiments. The listeners were divided into two groups, namely, trained and control groups. Each group consisted of 10 listeners.

A total of five tasks were performed. Four of these were identical to the tasks used in Experiment 1, and the fifth one was a newly employed ITD–high task. The ITD-high task was the same as the ITD task, except that a transposed stimulus (a 4-kHz tone, amplitude-modulated with a half-wave rectified 125-Hz tone; Bernstein and Trahiotis 2002) was used instead of the SSMC.

The listeners in both groups participated in 4 days of test sessions for all the five tasks. The test sessions on the first day (S0) were intended to familiarize the listeners with the general procedures and the stimuli used in the experiment. The second day (S1) was in the same week as S0, but the days were not consecutive. For the trained group, a 2-week training period (described later) started in the week after S1. The control group experienced the same number of days as the training sessions, but no psychophysical measurements or training was performed. The third and fourth days of the test sessions (S2 and S3, respectively) were 1 week and 1 month, respectively, after the last training session day. On each test session day, the five tasks were conducted in a random order. For each day and task, three threshold values were obtained by repeating the adaptive tracking, and the average of the three values was taken as that day’s threshold for the task. The thresholds on S1, S2, and S3 were regarded as pre-training, 1-week post-training, and 1-month post-training thresholds and were subject to data analysis.

For the trained group, the listeners underwent training in the pitch task, in which adaptive tracking was repeated 12 times per day. The 2-week training period included weekends and so were 12-day training sessions.

Figure 11.2a compares post-training thresholds with pre-training values for the pitch task (trained task), for the trained group of listeners. The symbols in the plot are generally below the diagonal line, indicating that the thresholds became lower after the training (i.e., performance improved). There was no such trend for the control group (data not shown). A repeated measures analysis of variance (ANOVA) on the thresholds (in a logarithmic scale) was conducted with factors of the test day (within-subject factor) and the listener group (between-subject factor). The main effects of the two factors were not significant (test day F(2,36)  =  1.38, p  =  0.2651; group F(1,36)  =  2.32, p  =  0.1454), but there was a marginal interaction of the test day and the listener group (F(2,36)  =  3.25, p  =  0.0505). When focusing on the trained group, an ANOVA indicated a significant effect of the test day (F(2,18)  =  4.25, p  =  0.0309). The difference between post- and pre-training thresholds was marginally significant for the 1-week post-training data (p  =  0.073, paired t-test) and was significant for the 1-month post-training data (p  =  0.032). A closer examination of the plot revealed that the degree of improvement tended to be smaller in listeners with lower pre-training thresholds. This can be explained as a floor effect, which was why the statistical test showed only a marginal significance for the 1-week post-training data. An ANOVA on the control group showed no effect of the test day (F(2,18)  =  0.77, p  =  0.4763).

The thresholds for the ITD task (untrained task) are shown in Fig. 11.2b. The 1-week post-training thresholds were not significantly different from the pre-­training thresholds (p  =  0.174), although there was a slight tendency for listeners with higher pre-training thresholds to exhibit elevated post-training thresholds. Thresholds obtained 1 month after the training session exhibited significant increases relative to the pre-training data (p  =  0.0005). A repeated measure ANOVA with factors of the test day and the listener group indicated a significant main effect of the test day (F(2,36)  =  3.58, p  =  0.0382), but not for the listener group (F(1,36)  =  0.34, p  =  0.5697). There was a significant interaction of the test day and the listener group (F(2,36)  =  3.95, p  =  0.0280). Separate ANOVAs indicated a significant effect of the test day for the trained group (F(2,18)  =  7.26, p  =  0.0049), but not for the control group (F(2,18)  =  0.03, p  =  0.9715).