This study will employ a randomized, cluster-organized, crossover design to compare the effects of a music-based intervention (TEST) with an audiobook listening intervention (CONTROL). Both interventions will be delivered by a therapist in the participants' living environment over ten evening sessions per intervention period. Each participant will serve as their own control and will be exposed to both intervention conditions. The sequence of intervention periods ("TEST followed by CONTROL" or "CONTROL followed by TEST") will be randomized at the cluster level within participating protected care units (UVP). Behavioral and actigraphic outcomes will be collected and analyzed at baseline, during, and after each intervention period.
