The primary purpose of the study is to evaluate whether the effect of bio-feedback dysphagia treatment is more effective than a control treatment in post-stroke patients.

This study will be a randomized, single-blind controlled pilot study. The investigators intend to recruit 40 patients who have suffered a stroke and have dysphagia. Twenty patients will undergo training with bio-feedback (experimental group) and the other 20 patients will undergo standard training, using only verbal feedback rather than visual feedback (control group).

The authors hypothesize that in the control group the efficacy of the treatment will be lower in the absence of immediate visual feedback of strength and coordination of the swallowing act.

The procedure is divided into 5 parts:

• Clinical and instrumental evaluation (fibroendoscopy) of swallowing functions before the treatment;
• 25 sessions with bio feedback treatment or 25 sessions of control treatment;
• Clinical and instrumental evaluation (fibroendoscopy) of swallowing functions after the training;
• Clinical and instrumental evaluation (fibroendoscopy) of swallowing functions after 3 months.

Each participant will attend 5 sessions per week for a total duration of 5 weeks. The duration of each session will take 1 hour, including bio-feedback preparation that will take about 10 minutes for assembly and 5 minutes for removal. At the beginning of the bio-feedback session, a surface electrode will be applied to the mylohyoid muscle. The computer will analyze the electromyographic signal of the patient in real time and will produce visual feedback of the muscular activity on the screen. The patient will be required to swallow at a given time, with food bolus if possible, and perform maneuvers that favor swallowing strength and efficacy.

During the first session, participant will be instructed to use the tool before starting the training phase. Patients of both groups will participate in similar sessions, with classical maneuvers favoring swallowing efficacy that is effortful swallow, supraglottic swallow, and "Masako" maneuver, the first two with bolus administration, if possible. The experimental group will perform this training for 45 minutes with the support of visualization of muscle activity on the screen (biofeedback), while the control group will attend the same training for 45 minutes, but receiving verbal feedback from the speech and language therapist.

Data analysis

On the measures collected, a descriptive statistical analysis and distribution will be studied. Initial instrumental and clinical data from FOIS scale (The Functional Oral Intake Scale), PAS (Penetration-Aspiration Scale), Pooling Score (Farneti et al.); Cranial nerves test will be analyzed using parametric (t-test) and non-parametric (Wilcoxon test) tests with a p-value considered significant if <0.05, while longitudinal variations will be analyzed using ANOVA for repeated measurements. In addition, the effect size (Cohen's d) of the difference between the instrumental and clinical evaluation scales average before and after training in both groups will be calculated. This will allow to have an estimate of training effectiveness to perform a power study and plan a wider experimental study for validating the method.

Inclusion Criteria:

• first stroke injury
• single lesion
• onset from more than six weeks
• presence of dysphagia (Dysphagia Outcome and Severity Scale ≥ 4)
• good comprehension ( not below PT 53 in Token Test)
• good eyesight and hearing, or adequately corrected

Exclusion Criteria:

• All the elements that hinder a good recording of the electromyographic signal (for instance a cutaneous infection, a wound or a dermatitis in the submandibular area);
• Severe concomitant illnesses (fever, infections, metabolic problems, serious cardiac insufficiency) that can influence patient's collaboration;
• Serious dystonia or unintentional movements
• presence of disorders of consciousness;
• encephalopathy due to multiple infarcts;
• inability to sustain the experimental treatment;
• other neurological diseases (for example Parkinson disease);