McKechnie Jacqueline, Shahin Mostafa, Ahmed Beena, McCabe Patricia, Arciuli Joanne, Ballard Kirrie J
Faculty of Health Sciences, The University of Sydney, Camperdown, NSW 2006, Australia.
Faculty of Health, University of Canberra, Bruce, ACT 2617, Australia.
Brain Sci. 2021 Oct 25;11(11):1408. doi: 10.3390/brainsci11111408.
Childhood apraxia of speech (CAS) commonly affects the production of lexical stress contrast in polysyllabic words. Automated classification tools have the potential to increase reliability and efficiency in measuring lexical stress. Here, factors affecting the accuracy of a custom-built deep neural network (DNN)-based classification tool are evaluated. Sixteen children with typical development (TD) and 26 with CAS produced 50 polysyllabic words. Words with strong-weak (SW, e.g., nosaur) or WS (e.g., bana) stress were fed to the classification tool, and the accuracy measured (a) against expert judgment, (b) for speaker group, and (c) with/without prior knowledge of phonemic errors in the sample. The influence of segmental features and participant factors on tool accuracy was analysed. Linear mixed modelling showed significant interaction between group and stress type, surviving adjustment for age and CAS severity. For TD, agreement for SW and WS words was >80%, but CAS speech was higher for SW (>80%) than WS (~60%). Prior knowledge of segmental errors conferred no clear advantage. Automatic lexical stress classification shows promise for identifying errors in children's speech at diagnosis or with treatment-related change, but accuracy for WS words in apraxic speech needs improvement. Further training of algorithms using larger sets of labelled data containing impaired speech and WS words may increase accuracy.
儿童言语失用症（CAS）通常会影响多音节词中词汇重音对比的产生。自动化分类工具具有提高测量词汇重音的可靠性和效率的潜力。在此，对影响基于定制深度神经网络（DNN）的分类工具准确性的因素进行评估。16名发育正常（TD）儿童和26名患有CAS的儿童说出了50个多音节词。具有强弱（SW，例如，恐龙）或弱强（WS，例如，香蕉）重音的词被输入到分类工具中，并针对以下情况测量准确性：（a）与专家判断对比，（b）按说话者群体，以及（c）样本中是否有音素错误的先验知识。分析了片段特征和参与者因素对工具准确性的影响。线性混合模型显示群体和重音类型之间存在显著交互作用，在对年龄和CAS严重程度进行调整后仍然显著。对于TD儿童，SW和WS词的一致性>80%，但CAS儿童的言语中SW词（>80%）的一致性高于WS词（约60%）。音段错误的先验知识并没有带来明显优势。自动词汇重音分类在识别儿童诊断时或与治疗相关变化时的言语错误方面显示出前景，但失用性言语中WS词的准确性需要提高。使用包含受损言语和WS词的更大标记数据集对算法进行进一步训练可能会提高准确性。
