Educational equality and the reduction of discrimination are among the UN’s Sustainability Goals. Previous studies as well as policy recommendations suggest that the extent to which these are implemented in the field of speech and language therapy for multilingual children depends on sufficient knowledge and material. To this end, an online survey was carried out with 300 Speech and Language Therapists (SLTs) from Austria, Germany, Italy, and Switzerland, investigating their attitudes and approaches regarding the service provision for multilingual children. Their responses were analyzed taking the SLTs’ language background, experience, and country of origin into account. Results were interpreted in the context of country-specific SLT service-related policies and SLT training as well as migration history. There seems to be a gap between the SLTs’ knowledge about the specific requirements for providing Speech Language Therapy (SLT) for multilingual children and their common practice, which—despite the continuous need of further training—points to sufficient awareness but a lack of materials or resources. We found experience in working with multilingual children to be the most influential factor on attitudes and approaches towards multilingualism. This suggests the importance of improving pre-exam and early-career professional experience to foster SLTs’ development of mindful attitudes and appropriate approaches towards multilingualism in their clinical practice.

No abstract

No abstract

This paper presents a fully automated approach for identifying speech anomalies from voice recordings to aid in the assessment of speech impairments. By combining Connectionist Temporal Classification (CTC) and encoder-decoder-based automatic speech recognition models, we generate rich acoustic and clean transcripts. We then apply several natural language processing methods to extract features from these transcripts to produce prototypes of healthy speech. Basic distance measures from these prototypes serve as input features for standard machine learning classifiers, yielding human-level accuracy for the distinction between recordings of people with aphasia and a healthy control group. Furthermore, the most frequently occurring aphasia types can be distinguished with 90% accuracy. The pipeline is directly applicable to other diseases and languages, showing promise for robustly extracting diagnostic speech biomarkers.