Back to Search Start Over

Communication improvement reduces BPSD: a music therapy study based on artificial neural networks.

Authors :
Raglio, Alfredo
Bellandi, Daniele
Manzoni, Luca
Grossi, Enzo
Source :
Neurological Sciences. May2021, Vol. 42 Issue 5, p2103-2106. 4p. 1 Graph.
Publication Year :
2021

Abstract

Background: Literature showed the effects of music therapy on behavioral disturbances, cognitive functions, and on quality of life in people with dementia. Especially, relational active music therapy approach is oriented to reduce behavioral disturbances increasing communication, especially non-verbal communication. Objective: This study aimed at exploring the connection between the baseline characteristics of responders and the positive outcome of the intervention, but also the close relationship between the behavioral disturbances and the core of the therapeutic intervention (the relationship/communication improvement). Method: Linear correlation index between input variables and the presence of a critical improvement of behavioral symptoms according Neuropsychiatric Inventory and a semantic connectivity map were used to determine, respectively, variables predictive of the response and complex connections between clinical variables and the relational nature of active music therapy intervention. The dataset was composed of 27 variables and 70 patients with a moderate-severe stage of dementia and behavioral disturbances. Results: The main predictive factor is the Barthel Index, followed by NPI and some of its sub-items (mainly, Disinhibition, Depression, Hallucinations, Irritability, Aberrant Motor Activity, and Agitation). Moreover, the semantic map underlines how the improvement in communication/relationship is directly linked to "responder" variable. "Responder" variable is also connected to "age," "Mini Mental State Examination," and sex ("female"). Conclusions: The study confirms the appropriateness of active music therapy in the reduction of behavioral disturbances and also highlights how unsupervised artificial neural networks models can support clinical practice in defining predictive factors and exploring the correlation between characteristics of therapeutic-rehabilitative interventions and related outcomes. [ABSTRACT FROM AUTHOR]

Details

Language :
English
ISSN :
15901874
Volume :
42
Issue :
5
Database :
Academic Search Index
Journal :
Neurological Sciences
Publication Type :
Academic Journal
Accession number :
149788807
Full Text :
https://doi.org/10.1007/s10072-020-04986-2