| Classical threshold theory assumes that a threshold is fixed and immutable. It cannot explain nor predict the underlying mechanisms for any changes in threshold value due to training. However, increased sensitivity to sensory signals through learning has been documented. This perceptual learning has been found to occur in such sensory tasks such as visual, auditory, and somatosensory signal detection/discrimination. Our experiments examine the degree of specificity and various features of perceptual learning using psychophysical measures. Based on our data we can make inferences about the neuronal mechanism underlying this type of learning. Experiments 1, 2, and 3 demonstrated that learning is specific for the trained frequency, the temporal order of tone presentation, and the noise context of the auditory stimuli. The results are consistent with the view that perceptual learning cannot be fully explained by the reorganization of the tonotopic map in the auditory cortex. Learning does not lead to a general increase in sensitivity to the trained tones. Experiment 4 confirmed that external feedback is not necessary for perceptual learning to occur. We hypothesized that internal feedback may exist to allow effective learning. One possible source of internal feedback is the observer's ability to evaluate their own performance during trials in which the tones were distinguishable. We speculated that discrimination threshold might not be stationary. Sensory stimuli in discrimination trials with sub-threshold stimuli might be clearly differentiated on some trials, and these trials can provide the internal feedback for inducing effective perceptual learning. Experiment 5 explored these hypotheses by examining the possibility that auditory thresholds are non-stationary (i.e., they vary over time). The results provide strong evidence of non-stationary thresholds for auditory frequency discrimination. Altogether, the data (1) demonstrate that sensory thresholds can fluctuate over time, that (2) these fluctuations may support the surprising ability of listeners to improve their performance without external guidance, and that (3) perceptual learning is very specific to the exact conditions in which the learning occurred. This specificity is inconsistent with any model that suggests that perceptual learning of frequency discriminations entails any generalized increase in the acuity of auditory percepts. |
