Generalization to untrained conditions following training with identical stimuli

It has recently been documented that perceptual learning occurs in the absence of perceived differences between stimuli. This finding supports the notion that stimuli do not have to be highly discriminable in order to elicit learning effects. It is not clear, however, whether additional characteristics of perceptual learning, such as generalization of the learning gains, will be demonstrated in this type of training. Thus, the goals of the present study were to compare the effects of type of training, i.e. adaptive versus fixed with identical stimuli on (1) difference limen for frequency (DLF) thresholds; and (2) the extent of generalization of the learning gains to an untrained frequency and the untrained ear. Twelve adult participants with normal hearing were randomly assigned into two experimental groups: an 'adaptive' group and a 'fixed' group. A single-session frequency discrimination training/testing protocol was conducted in which participants were assessed using three 1,000 Hz DLF testing intervals separated by two training intervals. In the 'adaptive' group, each training interval consisted of tone pairs that varied in a 2-down 1-up adaptive procedure. In the 'fixed' group, each training interval consisted of identical 1,000 Hz tone pairs. Following the training/testing protocol, two untrained conditions were tested: a 2,000 Hz untrained frequency and the untrained ear. Results show that: (1) training on a DLF task with identical stimuli improved performance; (2) the type of training (i.e., 'adaptive' versus 'fixed' with identical stimuli) did not influence the amount of gains in performance; (3) learning generalized to an untrained frequency but to a lesser extent in the 'fixed' group with identical stimuli compared to the 'adaptive' group; and (4) learning generalized to the untrained ear for both groups. Thus, training with identical stimuli appears to promote auditory perceptual learning in a frequency discrimination task that generalizes, to some extent, to untrained conditions. These findings support the hypothesis that top-down attention processes and not only bottom-up processes may play an important role in auditory perceptual learning.