Efficacy of Ultrasound Biofeedback in Brazilian Childhood Apraxia of Speech: A Randomized Clinical Trial

Ultrasound Biofeedback Intervention (Experimental Group): The experimental intervention involves intensive speech therapy augmented with real-time ultrasound tongue imaging as visual feedback. A portable ultrasound system with a small microconvex transducer is used, positioned submentally (under the child's chin) to capture a midsagittal view of the tongue.

. The ultrasound transducer emits safe, high-frequency sound waves, and the reflected signals produce a dynamic image of the tongue's surface on a monitor. Specialized software (such as Articulate Assistant Advanced (AAA) by Articulate Instruments Ltd.) is utilized to display and synchronize the ultrasound images with speech output.

During therapy, the child can see a real-time outline of their tongue moving on the screen as they attempt target sounds. This visual biofeedback serves as knowledge of performance - it allows the child to compare their tongue shape and placement against the therapist's model or a predetermined target shape. For example, the clinician may instruct the child to "lift the back of your tongue" for a velar sound and the ultrasound image immediately shows whether the child's tongue dorsum is reaching the correct position. A custom head-stabilization apparatus is often used to keep the probe steady and ensure consistent imaging. Throughout the ultrasound-assisted practice, the therapist provides cues and coaching based on the visual information (e.g. telling the child to aim the tongue toward a certain position on the screen) in addition to the typical auditory and tactile cues.

Control Intervention (Conventional Motor-Based Therapy): The control group receives an intensive motor-based speech therapy program of equal dose and duration, but without ultrasound visual feedback. This conventional therapy is grounded in established motor learning principles for CAS, focusing on repetitive practice of speech targets with gradually increasing difficulty and strategic feedback. Children in the control group practice the same type of target sounds and words as the experimental group (described below under "Therapy Targets and Sessions"). The key difference is that the clinician uses standard cueing techniques instead of instrumented biofeedback: for example, verbal descriptions of tongue placement ("Put your tongue tip behind your teeth for /t/"), visual cues (watching the clinician's mouth, using a mirror), tactile prompts, and knowledge of results feedback (indicating whether the attempt was correct or not). Both groups therefore receive high-intensity, drill-based practice and employ principles of motor learning (such as frequent practice, feedback, and gradually reduced cueing), but the experimental group gains extra visual feedback about tongue movements, whereas the control relies on traditional auditory-proprioceptive feedback only. This distinction reflects different motor learning approaches: the ultrasound provides immediate internal feedback on articulator movement, while the control approach emphasizes outcomes (sound correctness) without direct internal visual feedback.

Therapy Targets and Session Structure: Therapy targets are individualized for each child based on their speech error profile, focusing on phonemes and word shapes that are challenging due to CAS. Targets include specific phonemes (speech sounds) and various syllable structures (e.g., CV syllables, CVC words, consonant clusters, and multisyllabic words) that the child has difficulty producing. For instance, a child who cannot accurately produce lingual consonants might have targets like /s/, /ʃ/, or /k/ in simple syllables ("sa", "ka") and in words of increasing complexity ("sapato", "cavalo", etc.). Both groups practice a similar set of target stimuli (words or pseudowords containing these phonemes in different syllabic contexts), chosen to facilitate motor planning and to allow measurement of generalization to unpracticed words. Each participant undergoes a series of therapy sessions over the course of the intervention. Sessions are typically scheduled multiple times per week (e.g. 2 sessions per week) for several weeks, to accumulate a high dose of practice trials. Each session lasts about 45-60 minutes, structured into two phases: an ultrasound-assisted phase and an unassisted phase. In the experimental group's sessions, roughly the first half of the session is the ultrasound-assisted practice. During this phase (approximately 20-30 minutes), the child practices the target sounds/words while viewing the real-time ultrasound display of their tongue. The clinician might use 8-10 target words per session, training each multiple times. The child is encouraged to adjust their articulations by watching the tongue image (for example, ensuring the tongue shape reaches a certain height or contact for a sound) and by following the clinician's live feedback on the screen. In the second half of the session, the ultrasound machine is turned off (or the screen is hidden) for the unassisted practice phase, which also lasts about 20-30 minutes.