This study investigated whether different types of dysarthric speech (with comparable baseline intelligibility) are equally susceptible to background noise. Using intrinsically degraded speech from four individuals with dysarthria, each representing a distinct motor speech disorder subtype, we examined how intelligibility is impacted by masker type (stationary vs fluctuating noise) and signal-to-noise ratio (SNR). Ninety-five listeners completed a listening task with stationary speech-shaped noise or ecologically valid cafeteria noise, across multiple SNRs. Although all degraded signals showed intelligibility declines in noise, the extent and pattern of decline varied, suggesting that not all forms of pathological degradation interact with background noise in the same way. Notably, the typical intelligibility benefit associated with fluctuating maskers (i.e., masking release) was absent for all dysarthric signals. These findings reveal that equivalently intelligible but qualitatively different dysarthric signals do not respond uniformly to environmental challenges. The results have implications for models of speech perception, which must account for complex interactions between intrinsic (speaker-specific) and external (environmental) degradations, as well as for real-world communication and intervention strategies.

Automated analysis of speech and language provides a critical opportunity for developing a scalable tool for general medical settings to aid in psychiatric diagnosis and triage. In a cross-diagnostic highly comorbid sample, we evaluated the contribution of speech and language features to diagnostic classification through a hierarchical, dichotomous approach: first distinguishing healthy volunteers (HV) vs. participants with any psychiatric disorder (Split 1), then identifying those with serious mental illness (SMI) vs. other psychiatric disorders (Split 2). Speech was collected from 266 participants via picture description, verbal fluency, paragraph reading, and open-ended verbal journaling tasks. We extracted 640 interpretable features spanning acoustic, temporal, lexical, syntactic, discourse, and coherence domains. LightGBM models were trained with 5-fold cross-validation comparing all combinations of speech tasks, and SHAP values were plotted for feature importance. For Split 1, the best-performing model achieved F1=0.865 combining picture description and journaling tasks, with picture description alone reaching F1=0.830. For Split 2, performance was moderate (F1=0.626), with picture description also as the best single task. Top features for Split 1 included amplitude instability (shimmer) and restricted pitch variance; Split 2 was influenced by articulation rate and filled pauses. This study represents a novel classification approach in a naturalistic, clinically complex sample. Findings suggest that a brief, explainable speech-based assessment may be able to identify individuals who need further evaluation for psychiatric disorders. External validation, bias auditing, and deployment studies are warranted to assess clinical impact.

PurposeChildren with cleft conditions are at increased risk of language and speech disorders. Velopharyngeal dysfunction is a common sequela of palate repair, and children may develop cleft-type speech characteristics, with consequences in the academic, social, and ultimately employment, domains. Infants with certain speech behaviors are more likely to present with increased intelligibility at school age. These behaviors could be embedded into general language development strategies and taught to infants by their caregivers. This paper describes a study protocol investigating the effects of early language intervention strategies, as part of a caregiver-coached, group-based intervention in infants with cleft palate. The goal will be measured by changes in comprehension and expression, as well as the ratio of word to gesture use, attempts at oral plosive sounds, number of phonemes, and frequency and shape of babbling.MethodsWe will recruit up to 40 infants with cleft palate, plus their caregivers. Infants should be aged between 6 and 15 months. Families will attend a total of 12 sessions over 6 weeks and will complete demographic and communication-based questionnaires. Caregiver confidence will also be measured.ResultsWe hypothesize that, in addition to increasing the number of words or concepts that the infant can comprehend and express, there will be an increase in the specific speech behaviors that will be targeted in this intervention.ConclusionIf the study hypotheses are confirmed, these findings will provide support to better equip families with strategies to support speech and language development in the first 15 months of life.

Preoperative speech deficits in children with posterior fossa tumours (PFTs) may be linked to postoperative speech impairment (POSI), a hallmark of posterior fossa syndrome, yet this relationship remains underexplored. POSI is characterized by mutism or markedly reduced speech that resolves within weeks to months but often leaves persistent motor-speech deficits. Most existing literature has focused on postoperative outcomes, with limited attention to preoperative speech function. We therefore systematically investigated this association by analyzing perceptible motor-speech features in a large cohort of patients scheduled for surgery within the European Study of Cerebellar Mutism Syndrome. We included data from 135 children which formed three age-and sex- comparable groups: 16 patients who later developed POSI, 11 were mute and five had severely reduced speech (3;9-14;4 years), 62 patients with habitual speech (3;2-17;11 years), and 57 healthy controls (3;0-18;1 years). Patients were Italian-, Dutch-, and English-speaking, while controls were Dutch- and Italian-speaking. Four speech-language therapists rated participants' narrative recordings on 23 speech features (e.g., distorted vowels, imprecise consonants, stuttering, and hoarseness). Group comparisons revealed significant preoperative differences between the children who later developed POSI and the children with habitual speech in both hypernasality and in phonation-respiration (e.g., leaky, hoarse, and strained voice). In addition, patients who later developed POSI showed impairment compared to healthy controls in prosody (e.g., speech rate, excess and equal stress) and articulation (e.g., distorted vowels, imprecise consonants, sequencing errors). Patients with habitual speech differed from the controls in hypernasality and prosody. Moreover, age influenced the severity of articulatory and prosodic symptoms, with younger patients showing greater impairment. Children with PFT show signs of speech impairment preoperatively relative to healthy controls. When comparing patients, those later developing POSI show more severe impairments related to resonance and voice quality. This suggests possible differences in the degree to which speech-related underlying anatomical substrates were compromised, and these early differences may represent potential predictors of later POSI. These findings highlight the value of preoperative assessment for identifying children at a heightened risk of POSI, opening opportunities for prehabilitation - targeted early intervention in the form of speech therapy - to strengthen the speech-motor networks.

Pediatric speech sound disorders (SSDs) affect many young children and are commonly assessed through auditory-perceptual judgments and IPA transcription, which are limited by listener bias, variable interrater reliability, and difficulty attributing deviations at the phoneme level. We present a developmentally informed framework for pediatric vocal biomarker foundations that prioritizes age-aware, phoneme-resolved interpretability over utterance-level accuracy alone.

Background/Objectives: Speech disorders caused by neurological, articulatory, cognitive, and behavioral impairments require accurate multimodal analysis frameworks capable of jointly understanding acoustic abnormalities and neurolinguistic inconsistencies for reliable computer-assisted speech disorder assessment. Recent multimodal deep learning approaches have utilized speech signals, linguistic transcripts, attention mechanisms, and contextual fusion strategies to improve disordered speech analysis and intelligent disorder prediction. However, existing methods frequently suffer from insufficient temporal synchronization, ineffective local-global feature learning, multimodal redundancy, poor interpretability, and reduced robustness under heterogeneous speech disorders. To address these challenges, Methods: The proposes NSX-Net (NeuroSpeech Explainable Network), an intelligent multimodal deep learning framework for speech disorder classification using Acoustic Speech Signal Data and Neurolinguistic Text/Linguistic Data. Initially, the Adaptive Speech Refinement Module (ASRM) performs noise removal, silence elimination, signal normalization, spectrogram generation, MFCC extraction, and transcript preprocessing to improve multimodal speech consistency. Subsequently, the Hierarchical Multimodal Feature Learning Unit (HMFLU) extracts discriminative feature representations through the Cross-Domain Representation Encoder (CDRE), Fine-Grained Speech Pattern Analyzer (FGSPA), and Global Sequential Dependency Learner (GSDL) for capturing both local articulation abnormalities and long-range semantic dependencies. Furthermore, the Dual-Path Attention Enhancement Block (DPAEB) emphasizes clinically important disordered speech regions using adaptive local-global attention mechanisms, while the Temporal Resolution Synchronization Module (TRSM) aligns rhythm-level and beat-level multimodal contextual structures to improve temporal consistency and suppress noise disturbances. Results: Experimental evaluation across six publicly available multimodal speech disorder datasets demonstrated that NSX-Net achieved superior performance, with 99.27% Accuracy, 99.11% Precision, 98.97% Recall, 99.02% F1-Score, and 99.41% AUC, significantly outperforming existing state-of-the-art frameworks. Conclusions: Finally, optimized multimodal contextual representations are forwarded into a Softmax classification layer for intelligent multiclass speech disorder prediction with improved interpretability and potential clinical applicability.

Background Atlantoaxial osteoarthritis (AAOA) is a degenerative disorder of the C1-C2 articulation that causes occipitocervical pain, restricted head rotation, and functional limitation. Treatment ranges from nonoperative care to image-guided intra-articular injection and surgical stabilization in selected patients with persistent symptoms or structural compromise. This study evaluated the clinical and radiographic outcomes of patients with AAOA managed through a defined institutional treatment protocol. Methods This single-center observational cohort study included 232 patients with AAOA treated at a tertiary spine center. The cohort included retrospectively identified patients treated from January 2020 to March 2021 and prospectively followed patients treated from April 2021 to December 2025. All patients underwent initial nonoperative treatment according to an institutional protocol. Patients with persistent symptoms despite conservative care were offered image-guided intra-articular C1-C2 injection, and patients with refractory pain or structural compromise underwent posterior C1-C2 fusion. For analysis, patients were categorized according to definitive treatment into a nonoperative cohort and a surgical cohort. The nonoperative cohort included patients who improved with conservative treatment alone or with injection-based treatment and did not undergo surgery. Clinical outcomes were assessed using the Visual Analog Scale (VAS) and Neck Disability Index (NDI). Radiographic variables included laterality, joint type, osteoarthritis grade, periarticular ossicles, joint destruction, subluxation, instability, and fusion status in operated patients. Results A total of 232 patients were included, with 200 patients in the nonoperative cohort and 32 patients in the surgical cohort. All patients received initial conservative treatment, and 32 patients progressed to C1-C2 fusion. The surgical cohort was significantly younger than the nonoperative cohort, with a mean age of 50.0 ± 17.5 years compared with 65.7 ± 9.9 years, respectively. Bilateral involvement and mixed joint disease were the most common radiographic patterns in both cohorts. The surgical cohort showed greater radiographic severity, with higher rates of Grades 3 and 4 disease, joint destruction, subluxation, and instability. In the nonoperative cohort, the mean VAS score improved from 6.23 ± 1.68 at baseline to 1.61 ± 1.04 at final follow-up, and the mean NDI improved from 0.32 ± 0.11 to 0.06 ± 0.05. In the surgical cohort, the mean VAS score improved from 7.41 ± 1.19 to 1.97 ± 1.45, and the mean NDI improved from 0.60 ± 0.14 to 0.11 ± 0.10. Radiographic fusion was documented in 30 of 32 operated patients. Conclusions AAOA is a clinically important cause of occipitocervical pain and disability. In this single-center cohort managed through a defined institutional treatment protocol, most patients remained in the nonoperative pathway and demonstrated clinical improvement, while posterior C1-C2 fusion was associated with meaningful improvement in selected patients with refractory symptoms or structural compromise. Further prospective studies using standardized imaging, treatment protocols, and follow-up schedules are needed to refine management pathways for this condition.

Acute disseminated encephalomyelitis (ADEM) is a rare, immune-mediated demyelinating disorder of the central nervous system that typically follows viral or bacterial infections. There are only two known case reports that describe an association with ADEM post rickettsial infection. A female in her 80s presented with progressive neurological decline, including dysarthria, dysphagia and hemiparesis, following recent treatment in Hong Kong for rickettsial infection. Initial CT brain imaging showed a right subcortical infarct. She deteriorated with encephalopathy and respiratory failure, requiring intubation. MRI revealed extensive brain and spinal demyelination. Infectious, autoimmune and paraneoplastic investigations were negative, while serology revealed markedly elevated spotted fever group antibodies. Treatment with high-dose corticosteroids and plasma exchange therapy resulted in significant neurological recovery. She was discharged after rehabilitation with a good recovery at 70%-80% of pre-morbid function. This case supports a possible post-infectious immune-mediated mechanism linking rickettsia and ADEM and underscores the importance of early recognition and treatment of ADEM.

Neurological disorders can alter motor control, cognition, and speech production in subtle ways that are difficult to detect using standard diagnostic tools. Because conventional imaging techniques such as CT and MRI often fail to reveal functional abnormalities, there is growing interest in non-invasive digital health approaches. This study investigates the potential of speech-based deep learning for detecting concussion-related dysfunction using structured speech tasks. Audio recordings from 225 high school and college athletes (approximately 200 retained after quality filtering) performing eight speech tasks were analyzed. The recordings were converted into Mel spectrograms and processed using a convolutional neural network (CNN), with data augmentation applied to improve model generalization. The results demonstrate strong classification performance, particularly for tasks requiring dynamic articulation and rhythmic control. To better interpret model behavior, the speech tasks were grouped into three functional categories and examined using gradient-weighted class activation mapping (Grad-CAM). Sentence and word reading tasks, as well as rapid syllable repetition, achieved the highest accuracy (up to 96%) and exhibited focused activation patterns associated with phonetic complexity and motor timing. In contrast, sustained vowel tasks produced lower performance and more diffuse model attention, likely due to their limited acoustic variability. These findings highlight the effectiveness of CNN-based analysis of dynamic speech features and emphasize the importance of task selection in speech-based digital diagnostics. Although the framework may inform future speech-based neurological assessment research, this study focuses on mild traumatic brain injury (concussion) as an initial use case.