EPP216 - Emotional Speech Markers of Psychiatric Disturbance in Huntington’s Disease

EPP216

Emotional Speech Markers of Psychiatric Disturbance in Huntington’s Disease

L. Chenain ^{1 2 3,*}, A. Fabre ^{1 4 5 6}, H. Titeux ^{1 2 5}, G. Morgado ⁶, K. Youssov ^{1 2 5 6}, C. Clavel ³, A.-C. Bachoud-Lévi ^{1 2 5 6}

¹Département d'Etudes Cognitives, École Normale Supérieure, Paris, ²Faculté de Santé, Université Paris Est Créteil, INSERM U955, Institut Mondor de Recherche Biomédicale, Créteil, ³INRIA, ⁴Faculté de Santé, Université Paris Est Créteil, INSERM U955, Institut Mondor de Recherche Biomédicale, Paris, ⁵Service de Neurologie, AP-HP, Hopital Henri Mondor-Albert Chenevier, ⁶Centre d'Investigation Clinique, INSERM, Créteil, France

 

Introduction: Psychiatric symptoms and difficulties in emotional expression are a significant aspect of Huntington’s Disease (HD). Both may be associated or interact, but these associations remain unexplored. To improve patient follow-up, we investigated how emotional expression related to psychiatric symptoms in HD. Speech provides acoustic markers of emotional and psychiatric states across diverse populations, and is readily deployable for digital monitoring to support between-visit follow-up and enhance patient care. 

Objectives: To this aim, we developed emoHD, the first HD emotional/psychiatric speech corpus, to investigate the link between emotional expression and psychiatric symptoms. 

Methods: We included 102 HD gene carriers and 35 healthy controls (HC) from the REPAIR and BIOHD cohorts. Psychiatric symptoms were assessed using the Problem Behaviors Assessment (PBA), assessing depression, irritability/aggressivity, apathy and obsessive/compulsive scores. Participants performed emotional and non-emotional speech tasks. Their recordings were annotated using three emotional descriptors: primary emotions for a basic emotional characterization; affective phenomena selected based on PBA statements to capture complex HD-associated emotional states; and activation for emotional intensity. We analyzed (1) the emotional descriptors’ relationships using Fisher's tests, (2) emotional expression differences between HD and HC through permuted t-tests, and (3) the associations between emotions and psychiatric symptoms though stepwise multilinear regressions, using the proportions of the emotional labels as predictors and adjusting for sex, age and clinical group status (HD Integrated Staging System 0-1, 2, 3, HC).

Results: HD patients showed reduced emotional expressiveness compared to HC, with higher neutral activation levels. Affective phenomena were more powerful in identifying emotional expression differences between HD and HC, with HD expressing more apathetic, confused, depressed, disoriented, frustrated and pessimistic speech. In contrast, they expressed less of the primary emotion anger, and the affective phenomena irritable and other. Expression of specific emotions was congruent with psychiatric symptoms (Table 1).

Table 1. Positive associations with the PBA 

	Primary emotions	Affective phenomena
Depression		anxious nervous
Apathy	neutral	pessimistic depressed apathetic hopeless
Irritability/aggressivity	sad	quick tempered frustrated
Obsessive/compulsive		fearful apathetic

Conclusions: Emotional speech is a promising marker for psychiatric disturbances in HD. This encourages future work to develop speech-based digital monitoring tools like speech emotion recognition models. This paves the way for a scalable, non-invasive remote assessment of emotional and psychiatric symptoms in HD, with personalized care strategies and timely intervention.

 

Disclosure of Interest: None Declared