Table 1.
The Success of the Predictions of the Three Models (Unlimited, Limited Capacity, and Reverse Hierarchy Theory) for Experiments I–IV of Study 1
Table 2.
The Success of the Predictions of the Three Models (Unlimited, Limited Capacity, and Reverse Hierarchy Theory) for Experiments I and II of Study 2
Figure 1.
Results of Study 1, Experiments I–IV
Left: results using phonologically different word pairs (blue). Right: results using phonologically similar pairs (red). (A–H) The dynamics of the adaptive threshold assessment as a function of trial number (averaged across subjects ± SEM, n = 10 for each of the eight conditions). The level of the signal was modified in relation to subject's performance (following a three down–one up adaptive procedure). Illustrations of diotic (thick curves) and dichotic (thin curves) thresholds, which are calculated as the mean of last five reversals (see Materials and Methods), are marked by dashed lines in (A). Thresholds are denoted in decibel SNR. Binaural benefits (vertical arrows in all panels) are calculated as the difference between the diotic and dichotic thresholds.
(A and B) Experiment I: the identification task with no binaural uncertainty (consistent binaural protocol).
(C and D) Experiment II: the semantic task with no binaural uncertainty (consistent binaural protocol.
(E and F) Experiment III: the identification task with binaural uncertainty (mixed binaural protocol).
(G and H) Experiment IV: the semantic task with binaural uncertainty (mixed binaural protocol).
(I and J) A summary of the average binaural benefits obtained in Experiments I–IV (filled shaded bars), and the benefits calculated by an ideal listener model (open bars; see Figure S2 and Text S1), for phonologically different (left, [I]) and phonologically similar (right, [J]) pairs.
Figure 2.
Results of Study 2, Experiments I
Left: results using phonologically different words (blue). Right: results using phonologically similar words (red).
(A–D) The dynamics of the adaptive threshold assessment as a function of trial number (averaged across subjects ± SEM, n = 25). Notations as in Figure 1. Vertical arrows denote binaural benefits. All measurements were done using the mixed binaural protocol. (A and B) An identification task using a set size of two words.
(C and D) An identification task using a set size of ten words.
(E and F) A summary of the average binaural benefits obtained in the experiment (filled shaded bars) and the benefits calculated by the ideal listener model (open bars; see Figure S3), for the set size 2 and set size 10 conditions.
Figure 3.
Results of Study 2, Experiments II
Left: results using phonologically different word pairs (blue). Right: results using phonologically similar pairs (red).
(A–H) The dynamics of the adaptive threshold assessment for identification of word pairs as a function of trial number (averaged across subjects ± SEM, n = 15). Notations as in Figure 1. Vertical arrows denote binaural benefits.
(A and B) The adaptive protocol converging to 80% correct identification with no uncertainty (i.e., using the consistent binaural protocol).
(C and D) The adaptive protocol converging to 80% correct identification with uncertainty (mixed binaural protocol).
(E and F) The adaptive protocol converging to 60% correct identification with no uncertainty (consistent binaural protocol).
(G and H) The adaptive protocol converging to 60% correct identification with uncertainty (mixed binaural protocol).
(I and J) A summary of the average binaural benefits obtained in the experiment (filled shaded bars), and the benefits calculated by an ideal listener model (open bars; see Figure S4).