The effects of stimulus variability on the perceptual learning of speech and non-speech stimuli

The Role of Talker Variability in Nonnative Phonetic Learning: A Systematic Review and Meta-Analysis

PURPOSE

High-variability phonetic training (HVPT) has been found to be effective on adult second language (L2) learning, but results are mixed in regards to the benefit of multiple talkers over single talker. This study provides a systematic review with meta-analysis to investigate the talker variability effect in nonnative phonetic learning and the factors moderating the effect.

METHOD

We collected studies with keyword search in major academic databases including EBSCO, ERIC, MEDLINE, ProQuest Dissertations & Theses, Elsevier, Scopus, Wiley Online Library, and Web of Science. We identified potential participant-, training-, and study-related moderators and conducted a random-effects model meta-analysis for each individual variable.

RESULTS

On the basis of 18 studies with a total of 549 participants, we obtained a small-level summary effect size (Hedges' g = 0.46, 95% confidence interval [CI; 0.08, 0.84]) for the immediate training outcomes, which was greatly reduced (g = -0.04, 95% CI [-0.46, 0.37]) after removal of outliers and correction for publication bias, whereas the effect size for immediate perceptual gains was nearly medium (g = 0.56, 95% CI [0.13, 1.00]) compared with the nonsignificant production gains. Critically, the summary effect sizes for generalizations to new talkers (g = 0.72, 95% CI [0.15, 1.29]) and for long-term retention (g = 1.09, 95% CI [0.39, 1.78]) were large. Moreover, the training program length and the talker presentation format were found to potentially moderate the immediate perceptual gains and generalization outcomes.

CONCLUSIONS

Our study presents the first meta-analysis on the role of talker variability in nonnative phonetic training, which demonstrates the heterogeneity and limitations of research on this topic. The results highlight the need for further investigation of the influential factors and underlying mechanisms for the presence or absence of talker variability effects. Supplemental Material https://doi.org/10.23641/asha.16959388.

Stimulus variability and task relevance modulate binding-learning

Classical theories of attention posit that integration of features into object representation (or feature binding) requires engagement of focused attention. Studies challenging this idea have demonstrated that feature binding can happen outside of the focus of attention for familiar objects, as well as for arbitrary color-orientation conjunctions. Detection performance for arbitrary feature conjunction improves with training, suggesting a potential role of perceptual learning mechanisms in the integration of features, a process called "binding-learning". In the present study, we investigate whether stimulus variability and task relevance, two critical determinants of visual perceptual learning, also modulate binding-learning. Transfer of learning in a visual search task to a pre-exposed color-orientation conjunction was assessed under conditions of varying stimulus variability and task relevance. We found transfer of learning for the pre-exposed feature conjunctions that were trained with high variability (Experiment 1). Transfer of learning was not observed when the conjunction was rendered task-irrelevant during training due to pop-out targets (Experiment 2). Our findings show that feature binding is determined by principles of perceptual learning, and they support the idea that functions traditionally attributed to goal-driven attention can be grounded in the learning of the statistical structure of the environment.

Effects of stimulus repetition and training schedule on the perceptual learning of time-compressed speech and its transfer

Perceptual learning can facilitate the recognition of hard-to-perceive (e.g., time-compressed or spectrally-degraded) speech. Although the learning induced by training with time-compressed speech is robust, previous findings suggest that intensive training yields learning that is partially specific to the items encountered during practice. Here, we asked whether three parameters of the training procedure - the overall number of training trials (training intensity), how these trials are distributed across sessions, and the number of semantically different items encountered during training (set size) - influence learning and transfer. Different groups of participants (69 normal-hearing young adults; nine to 11 participants/group) completed different training protocols (or served as an untrained control group) and tested on the recognition of time-compressed sentences taken from the training set (learning), new time-compressed sentences presented by the trained talker (semantic transfer), and time-compressed sentences taken from the training set but presented by a different talker (acoustic transfer). Compared to untrained listeners, all training protocols yielded both learning and transfer. More intense training resulted in greater item-specific learning and greater acoustic transfer than less intense training with the same number of training sessions. Training on a smaller set size (i.e., greater token repetition during training) also resulted in greater acoustic transfer, whereas distributing practice over a number of sessions improved semantic transfer. Together, these data suggest that whereas practice on a small set that results in stimulus repetition during training is not harmful for learning, distributed training can support transfer to new stimuli, perhaps because it provides multiple opportunities to consolidate learning.

The perceptual learning of time-compressed speech: A comparison of training protocols with different levels of difficulty

Speech perception can improve substantially with practice (perceptual learning) even in adults. Here we compared the effects of four training protocols that differed in whether and how task difficulty was changed during a training session, in terms of the gains attained and the ability to apply (transfer) these gains to previously un-encountered items (tokens) and to different talkers. Participants trained in judging the semantic plausibility of sentences presented as time-compressed speech and were tested on their ability to reproduce, in writing, the target sentences; trail-by-trial feedback was afforded in all training conditions. In two conditions task difficulty (low or high compression) was kept constant throughout the training session, whereas in the other two conditions task difficulty was changed in an adaptive manner (incrementally from easy to difficult, or using a staircase procedure). Compared to a control group (no training), all four protocols resulted in significant post-training improvement in the ability to reproduce the trained sentences accurately. However, training in the constant-high-compression protocol elicited the smallest gains in deciphering and reproducing trained items and in reproducing novel, untrained, items after training. Overall, these results suggest that training procedures that start off with relatively little signal distortion (“easy” items, not far removed from standard speech) may be advantageous compared to conditions wherein severe distortions are presented to participants from the very beginning of the training session.

Reduced posterior parietal cortex activation after training on a visual search task

Gaining experience on a cognitive task improves behavioral performance and is thought to enhance brain efficiency. Despite the body of literature already published on the effects of training on brain activation, less research has been carried out on visual search attention processes under well controlled conditions. Thirty-six healthy adults divided into trained and control groups completed a pre-post letter-based visual search task fMRI study in one day. Twelve letters were used as targets and ten as distractors. The trained group completed a training session (840 trials) with half the targets between scans. The effects of training were studied at the behavioral and brain levels by controlling for repetition effects using both between-subjects (trained vs. control groups) and within-subject (trained vs. untrained targets) controls. The trained participants reduced their response speed by 31% as a result of training, maintaining their accuracy scores, whereas the control group hardly changed. Neural results revealed that brain changes associated with visual search training were circumscribed to reduced activation in the posterior parietal cortex (PPC) when controlling for group, and they included inferior occipital areas when controlling for targets. The observed behavioral and brain changes are discussed in relation to automatic behavior development. The observed training-related decreases could be associated with increased neural efficiency in specific key regions for task performance.