Children with cochlear implants (CI) have difficulty perceiving pitch differences, which contributes to poor melodic perception. Aims/Methods. This study investigated how children with bilateral CIs (n = 23, M = 6,9 years) and children with normal hearing (NH) (n = 27, M = 7,2 years) were able to discriminate pitch intervals with a semitone as the smallest difference, and recognize melodies presented without lyrics in two conditions: as a melody and as a melody combined with pictures. An interview related to music was also conducted. Results. Children with bilateral CIs achieved very high scores on the pitch discrimination task. Children with NH surpassed the children with bilateral CIs at identifying songs without pictures (children with CIs 11%, NH 52%) and together with pictures (CIs 30%, NH 83%).  Measurement of effect size showed moderate to strong differences between children with CIs/NH. The interview findings showed positive attitudes towards music, with 77% of the children with CIs (versus 100% of children with NH) saying they enjoyed listening to music, had favourite artists and favourite songs. Conclusion. Children with CIs demonstrated a very good ability to discriminate pitch differences, but scored low on melody recognition tasks. The results confirm that melody recognition is demanding for children with CIs. In spite of the difficulty CIs have transmitting musical information, these children with CIs as a group have developed positive attitudes towards music. This finding suggests the potential capacity of CIs in relationship to music, but even more the potential capacity of growing children’s brains.     Barn med cochleaimplantat (CI) har svårighet att uppfatta skillnader i tonhöjd, vilket påverkar förmågan att identifiera melodier. Syfte/metod. Syftet var att studera hur barn med bilaterala CI (n = 23, M = 6,9 år) kunde diskriminera mellan tonhöjdsintervall med en semiton som minsta enhet och identifiera fem barnsångsmelodier jämfört med barn med normal hörsel (NH), (n = 27, M = 7,2 år). Melodierna presenterades i två betingelser: utan stöd av bilder (UB) samt med stöd av bilder (MB). Dessutom genomfördes en intervju om musik. Resultat. Förmågan att diskriminera tonhöjdsskillnader var mycket god. Melodier (UB): barn med CI klarade 11 %, barn med NH 52 %. Melodier (MB): barn med CI klarade 30 %, barn med NH 83 %. Analys av effektstyrka mätt i Cohens d visade moderata till stora skillnader mellan barn med CI/NH. Deskriptiv analys av intervjuerna visade positiva attityder till musik och att 77 % av barn med CI (100 % av barn med NH) tyckte om att lyssna på musik, och att flera barn hade favoritartister och favoritlåtar. Slutsats. Barn med CI hade mycket god förmåga att uppfatta tonhöjdsskillnader, men hade svårt att identifiera melodier. Resultatet bekräftar att melodiigenkänning är krävande för barn med CI. Trots de begränsningar som CI har beträffande musikåtergivning har barnen som grupp positiva attityder till musik. Det torde säga en del om den potential CI har i relation till musik, men ännu mer om den stora kapacitet som uppväxande barns hjärnor har.

Similarity computation is a difficult issue in music information retrieval tasks, because it tries to emulate the special ability that humans show for pattern recognition in general, and particularly in the presence of noisy data. A number of works have addressed the problem of what is the best representation for symbolic music in this context. The tree representation, using rhythm for defining the tree structure and pitch information for leaf and node labelling has proven to be effective in melodic similarity computation. In this dissertation we try to built a system that allowed to classify and generate melodies using the information from the tree encoding, capturing the inherent dependencies which are inside this kind of structure, and improving the current methods in terms of accuracy and running time. In this way, we try to find more efficient methods that is key to use the tree structure in large datasets. First, we study the possibilities of the tree edit similarity to classify melodies using a new approach for estimate the weights of the edit operations. Once the possibilities of the cited approach are studied, an alternative approach is used. For that a grammatical inference approach is used to infer tree languages. The inference of these languages give us the possibility to use them to classify new trees (melodies).