Fig 1.
Discrimination and memory tasks.
In all tasks tones (horizontal bars) were presented sequentially. The discrimination tasks required listeners to pick the different (target; marked by an arrow) stimulus from two identical (standard) stimuli. The target differed from the standards in either frequency (FD; a, b) or duration (DD; c, d) and the standard stimulus frequency was either fixed (a, FDf; c, DDf) or roved (b, FDr; d, DDr) across trials. (e, f) The Tone n-back task was used as a measure of auditory working memory. Participants compared the current tone frequency with that n positions back.
Table 1.
Demographic information for participant groups.
Fig 2.
Correlations between FD and WM tasks in naïve participants.
Scatterplots show the correlations between pretest measures for FDf, FDr, Tone 3-back, and backward Digit Span tasks. Sample sizes (N) are provided in each panel with correlation statistics.
Fig 3.
FDr but not FDf training transferred to WM.
(a) FDf performance during training (N = 36) for individuals (grey lines) and group average (filled squares). (b) FDr performance during training (N = 24). (c) Improvement of FD trained and control groups on the Tone 3-back task, calculated as the improvement ratio of d’ [(posttest–pretest)/pretest]. (d) Correlation between WM improvement and FDr–FDf improvement (see text). Error bars here and in Figs 4–6 are ± 1 s.e.m.
Fig 4.
WM training and transfer to FD.
(a) Multiple-session training on the Tone n-back task (N = 17), with n first fixed at 2 (light blue lines) and then 3 (dark blue lines). (b) Pre- and posttest d’ for Tone 2- and 3-back. Only twelve of the control group (N = 36) were tested on the 2-back condition. (c) Transfer to FDf. Plotted are individual (grey circles) and group mean (filled squares) posttest thresholds, adjusted for differences in pretest thresholds. The horizontal dotted line and the grey patch indicate mean pretest threshold and 95% confidence interval. (d) Transfer to FDr. (e) Correlation between WM improvement and reduction of the influence of variability (FDr–FDf) in the trained participants.
Fig 5.
Transfer from Tone n-back to other WM tasks.
Plotted are individual (grey lines) and group mean (filled squares) pre- and post-training performance of the trained participants (N = 17; blue) and controls (N = 12; black) on (a) backward Digit Span, (b) digit 3-back, and (c) shape 3-back.
Fig 6.
Transfer from Tone n-back to tone duration discrimination.
(a, b) Correlation of DDf and DDr thresholds with Tone 3-back performance before training. (c, d) Individual (grey circles) and group mean (filled squares) posttest thresholds of the trained participants (N = 11) and controls (N = 15) on DDf and DDr, adjusted for differences in pretest thresholds. The horizontal dotted line and the grey patch indicate mean pretest threshold and 95% confidence interval. (e) Correlation of Tone 3-back and DDr improvements.
Fig 7.
DDr training and transfer to WM.
Plotted are individual (grey lines or circles) and group mean (filled squares) performance of the trained participants (N = 9) and controls (N = 15) (a) during training, (b) after training on DDr, adjusted for differences in pretest thresholds, and (c) before and after training on Tone 3-back.