Ultrax2020: Ultrasound Technology for Optimising the Treatment of Speech Disorders.

言语声音障碍(SSD)是儿童时期最常见的沟通障碍；16.5%的8岁儿童患有SSD，从只有两种语音中的一种到连家人都难以理解的言语。SSD可以单独发生，也可以是唐氏综合症、自闭症或腭裂等残疾的一部分。2015年，詹姆斯·林德联盟将提高沟通技能和调查干预方向确定为残疾儿童的两大研究重点。我们的研究计划旨在通过开发有助于评估、诊断和治疗SSD的技术来满足这一需求。目前在言语和语言治疗中，技术支持是稀少的。通过我们之前在Ultrax项目中的工作，我们展示了通过使用超声波对舌头进行实时成像，儿童可以快速学习发出他们以前似乎不可能发出的语音。通过这个项目，我们开发了增强舌头超声图像的技术，使其更清晰、更容易解释。Ultrax2020旨在推动这项工作，进一步将超声波舌头跟踪器开发成一种工具，用于诊断特定类型的SSD，并评估在NHS诊所使用超声波的容易程度。我们研究的最终目标是，言语和语言治疗师(SLT)将使用Ultrax2020自动评估和诊断SSD，从而实现更快、更有针对性的干预。正常情况下，言语评估包括倾听儿童并记录他们所说的话。这种方法可能会遗漏儿童说话方式中的重要微妙之处。例如，一个孩子可能会试着说“key”，但可能会被听到为“Team”。这导致SLT认为孩子无法区分t和k，并选择了一种旨在解决这一问题的疗法。然而，超声波可以让我们观察和测量舌头，揭示出在许多情况下，孩子们正在产生难以察觉的错误。在上面的例子中，放置在下巴下面的超声波扫描仪显示，孩子同时产生t和k。识别这些错误意味着SLT必须选择不同的治疗方法。然而，超声分析是一项耗时的任务，只有经过专业培训的语音科学家才能完成。开发一种自动分析超声的方法，从而创建一种更客观、更快使用的语音评估工具，是Ultrax2020的关键成果。在Ultrax项目工作的基础上，我们开发了一种跟踪舌头超声图像的方法，Ultrax2020旨在开发一种对舌形进行分类的方法，以形成自动评估的基础和一种客观衡量进展的方法。幸运的是，我们已经拥有了一个大型数据库，其中包括那些患有语言障碍的成年人和小学生的舌头运动的超声图像，以此为基础建立舌形分类模型并测试其性能。与此同时，我们将评估我们作为Ultrax2020的一部分开发的技术，方法是与NHS SLT合作，从患有各种SSD的儿童那里收集非常大的超声数据库。在三个不同的NHS诊所，SLT将记录100多名儿童在基于超声波的语言治疗前后的超声波。这些数据将被发送给大学言语科学家进行分析，并反馈给临床医生，建议干预方法。在项目接近尾声时，我们将能够将这种黄金标准的手工标记分析与项目期间开发的自动分类进行比较。在我们的研究项目结束时，我们将开发并验证一种新的超声评估和治疗工具(Ultrax2020)，供言语和语言治疗师用于SSD的诊断和治疗。

Speech Sound Disorders (SSDs) are the most common communication impairment in childhood; 16.5% of eight year olds have SSDs ranging from problems with only one of two speech sounds to speech that even family members struggle to understand. SSDs can occur in isolation or be part of disability such as Down syndrome, autism or cleft palate. In 2015, the James Lind Alliance identified improving communication skills and investigating the direction of interventions as the top two research priorities for children with disabilities. Our programme of research aims to fulfil this need by developing technology which will aid the assessment, diagnosis and treatment of SSDs. Currently in Speech and Language Therapy, technological support is sparse. Through our previous work in the Ultrax project we showed that by using ultrasound to image the tongue in real-time, children can rapidly learn to produce speech sounds which have previously seemed impossible for them. Through this project, we developed technology that enhances the ultrasound image of the tongue, making it clearer and easier to interpret. Ultrax2020 aims to take this work forward, by further developing the ultrasound tongue tracker into a tool for diagnosing specific types of SSDs and evaluating how easy it is to use ultrasound in NHS clinics. The ultimate goal of our research is that Ultrax2020 will be used by Speech and Language Therapists (SLTs) to assess and diagnose SSDs automatically, leading to quicker, more targeted intervention.Normally speech assessment involves listening to the child and writing down what they say. This approach can miss important subtleties in the way children speak. For example, a child may try to say "key" and it may be heard as "tea". This leads the SLT to believe the child cannot tell the difference between t and k and select a therapy designed to tackle this. However, ultrasound allows us to view and measure the tongue, revealing that in many cases children are producing imperceptible errors. In the above example, an ultrasound scanner placed under the chin shows that the child produces both t and k simultaneously. Identification of these errors means that the SLT must choose a different therapy approach. However, ultrasound analysis is a time consuming task which can only be carried out by a speech scientist with specialist training. It is a key output of Ultrax2020 to develop a method for analysing ultrasound automatically, therefore creating a speech assessment tool which is both more objective and quicker to use. Building on the work of the Ultrax project, where we developed a method of tracking ultrasound images of the tongue, Ultrax2020 aims to develop a method of classifying tongue shapes to form the basis of an automatic assessment and a way of measuring progress objectively. We are fortunate to already have a large database of ultrasound images of tongue movements from adults and primary school children, including those with speech disorders, on which to base the model of tongue shape classification and to test its performance. At the same time, we will evaluate the technology we develop as part of Ultrax2020 by partnering with NHS SLTs to collect a very large database of ultrasound from children with a wide variety of SSDs. In three different NHS clinics, SLTs will record ultrasound from over 100 children before and after ultrasound-based speech therapy. This data will be sent to a university speech scientist for analysis and feedback to clinicians recommending intervention approaches. Towards the end of the project, we will be able to compare this gold-standard hand-labelled analysis with the automatic classification developed during the project. At the conclusion of our research project we will have developed and validated a new ultrasound assessment and therapy tool (Ultrax2020) for Speech and Language Therapists to use in the diagnosis and treatment of SSDs.