Automated Speech Analysis for Psychosis Evaluation

FACUNDO CARRILLO; SIDARTA RIBEIRO; DIEGO F SLEZAK; MAURO COPELLI; GUILLERMO A CECCHI; NATALIA MOTA; MARIANO SIGMAN

Machine Learning and Interpretation in Neuroimaging

Springer

Año: 2016; p. 31 - 39

Psychosis is a mental syndrome associated to loss of con-tact with reality which may arise in patients with different diseases,such as schizophrenia or bipolar disorder. Symptoms include hallucina-tions, confused and disturbed thoughts or lack of self-awareness. Recentstudies have found that psychotic patients can be objectively screenedusing graph-theoretical algorithms for speech analysis. This analysisoften relies in manually executed tasks such as syntagma generation,text splitting or manual feature selection for classification. To solve thisfundamental limitation, we use three fully-automated text analysis toolsgraph generation methods. In addition, since aspects of psychosis may bemanifested in semantic aspects of speech, we also developed a semanticfeatures index based on speech coherence. We show that using this com-bined approach, classifications obtained from automatic techniques arehigher than 85 % in a database of 20 schizophrenic patients, with similarresults to previous works. In summary, here we develop and validate anew tool for automated speech processing which includes semantic andstructural aspects. The tool performs similar to manual screening pro-cedures providing a new method to complement standard psychometricscales and fostering automated psychiatric diagnosis.