Ecologically valid speech collection in behavioral research: The Ghent Semi-spontaneous Speech Paradigm (GSSP).

This paper introduces the Ghent Semi-spontaneous Speech Paradigm (GSSP), a new method for collecting unscripted speech data for affective-behavioral research in both experimental and real-world settings through the description of peer-rated pictures with a consistent affective load. The GSSP was designed to meet five criteria: (1) allow flexible speech recording durations, (2) provide a straightforward and non-interfering task, (3) allow for experimental control, (4) favor spontaneous speech for its prosodic richness, and (5) require minimal human interference to enable scalability. The validity of the GSSP was evaluated through an online task, in which this paradigm was implemented alongside a fixed-text read-aloud task. The results indicate that participants were able to describe images with an adequate duration, and acoustic analysis demonstrated a trend for most features in line with the targeted speech styles (i.e., unscripted spontaneous speech versus scripted read-aloud speech). A speech style classification model using acoustic features achieved a balanced accuracy of 83% on within-dataset validation, indicating separability between the GSSP and read-aloud speech task. Furthermore, when validating this model on an external dataset that contains interview and read-aloud speech, a balanced accuracy score of 70% is obtained, indicating an acoustic correspondence between the GSSP speech and spontaneous interviewee speech. The GSSP is of special interest for behavioral and speech researchers looking to capture spontaneous speech, both in longitudinal ambulatory behavioral studies and laboratory studies. To facilitate future research on speech styles, acoustics, and affective states, the task implementation code, the collected dataset, and analysis notebooks are available.

Speech as an indicator for psychosocial stress: A network analytic approach.

Recently, the possibilities of detecting psychosocial stress from speech have been discussed. Yet, there are mixed effects and a current lack of clarity in relations and directions for parameters derived from stressed speech. The aim of the current study is - in a controlled psychosocial stress induction experiment - to apply network modeling to (1) look into the unique associations between specific speech parameters, comparing speech networks containing fundamental frequency (F0), jitter, mean voiced segment length, and Harmonics-to-Noise Ratio (HNR) pre- and post-stress induction, and (2) examine how changes pre- versus post-stress induction (i.e., change network) in each of the parameters are related to changes in self-reported negative affect. Results show that the network of speech parameters is similar after versus before the stress induction, with a central role of HNR, which shows that the complex interplay and unique associations between each of the used speech parameters is not impacted by psychosocial stress (aim 1). Moreover, we found a change network (consisting of pre-post stress difference values) with changes in jitter being positively related to changes in self-reported negative affect (aim 2). These findings illustrate - for the first time in a well-controlled but ecologically valid setting - the complex relations between different speech parameters in the context of psychosocial stress. Longitudinal and experimental studies are required to further investigate these relationships and to test whether the identified paths in the networks are indicative of causal relationships.

BioLORD-2023: semantic textual representations fusing large language models and clinical knowledge graph insights.

OBJECTIVE: In this study, we investigate the potential of large language models (LLMs) to complement biomedical knowledge graphs in the training of semantic models for the biomedical and clinical domains. MATERIALS AND METHODS: Drawing on the wealth of the Unified Medical Language System knowledge graph and harnessing cutting-edge LLMs, we propose a new state-of-the-art approach for obtaining high-fidelity representations of biomedical concepts and sentences, consisting of 3 steps: an improved contrastive learning phase, a novel self-distillation phase, and a weight averaging phase. RESULTS: Through rigorous evaluations of diverse downstream tasks, we demonstrate consistent and substantial improvements over the previous state of the art for semantic textual similarity (STS), biomedical concept representation (BCR), and clinically named entity linking, across 15+ datasets. Besides our new state-of-the-art biomedical model for English, we also distill and release a multilingual model compatible with 50+ languages and finetuned on 7 European languages. DISCUSSION: Many clinical pipelines can benefit from our latest models. Our new multilingual model enables a range of languages to benefit from our advancements in biomedical semantic representation learning, opening a new avenue for bioinformatics researchers around the world. As a result, we hope to see BioLORD-2023 becoming a precious tool for future biomedical applications. CONCLUSION: In this article, we introduced BioLORD-2023, a state-of-the-art model for STS and BCR designed for the clinical domain.

Parents' perceptions on speech therapy delivery models in children with a cleft palate: A mixed methods study.

PURPOSE: This study investigated parents' perceptions on two different speech therapy delivery models in children with a CP ± L, namely an innovative high intensity speech intervention (i.e. HISI: 10 1-h sessions divided over 2 weeks) and a low intensity speech intervention (i.e. LISI: 10 1-h sessions divided over 10 weeks). METHOD: Twelve parents of 12 children who received HISI (n = 6) or LISI (n = 6) were contacted with the request to participate to this study to review their opinion on the received therapy. Participation included the completion of a questionnaire containing items related to satisfaction, speech progress, intervention intensity and frequency, transfer, and need for further speech therapy. Additionally, semi-structured interviews were carried out. The interviews were analyzed using an inductive thematic approach. RESULTS: There were no significant differences between the two groups in satisfaction with the "general speech therapy, "duration of one speech therapy session", "total intervention duration" and "degree of improvement of speech intelligibility". Following HISI, parents perceived more improvement in terms of spontaneous speech and better resolution of the speech disorders. The interviews revealed 3 themes of importance to the parents: (1) treatment-related expectations, (2) treatment-related burden, and (3) patient-therapist relationship. Parents in the HISI group reported two concerns: (1) the lack of variation when receiving daily intervention, and (2) the emotional burden when the child is confronted with his/her speech disorder on a daily basis. CONCLUSIONS: Parents were equally satisfied with the provided intervention. Parents in the HISI group perceived more speech progress following the intervention compared to parents in the LISI group. The intensive contact with the speech pathologist enhanced the patient-therapist relationship. To support a cultural shift away from low intensity therapy delivery models, it will be important to counsel and inform parents of the benefits of HISI and to counterbalance concerns.

Is High-Intensity Speech Intervention Better? A Comparison of High-Intensity Intervention Versus Low-Intensity Intervention in Children With a Cleft Palate.

Purpose The purpose of this study was to compare the effect of speech intervention provided with a low intensity with speech intervention provided with a high intensity on the speech and health-related quality of life (HRQoL) in Dutch-speaking children with a cleft palate with or without a cleft lip (CP ± L) between 4 and 12 years. Method A longitudinal, prospective, randomized controlled trial with a multiple baseline design was used. Twelve children with a CP ± L (M age = 8.0 years, SD = 1.54) were divided into two groups using block randomization stratified by age and gender: One group received low-intensity speech intervention (LISI; n = 6) and one group received high-intensity speech intervention (HISI; n = 6). Children in the LISI group received intervention with a session duration of 1 hr, a dose frequency of 1 session per week, and a total intervention duration of 10 weeks. Children in the HISI group received intervention with a session duration of 1 hr, a dose frequency of 5 sessions per week, and a total intervention duration of 2 weeks. The cumulative intervention intensity was kept constant. Both groups received identical therapy programs provided by the same experienced speech therapist. Perceptual speech assessments were performed on baseline and posttreatment data points. Changes in HRQoL were assessed using the Velopharyngeal Insufficiency Effects on Life Outcomes (VELO) questionnaire. Both groups were compared over time using (generalized) linear mixed models. Results No significant Time × Group interactions were observed for the percentage of correctly produced consonants at the word and sentence levels, indicating no differences in evolution over time among the two groups. The variables speech understandability, speech acceptability, and the total VELO scores significantly improved following HISI, but not following LISI. Conclusions Children in the HISI group made equal and, for some variables, even superior progress in only 2 weeks of therapy compared to children in the LISI group who received 10 weeks of therapy. HISI is a promising strategy to improve speech outcomes and HRQoL in a shorter time period.