Speech and language features as indicators of neuropsychiatric symptoms in a memory clinic population

Abstract

INTRODUCTION: Neuropsychiatric symptoms (NPSs) are early hallmarks of neurocognitive disorders (NCDs). Speech alterations might indicate both cognitive and behavioral changes in NCDs, aiding in diagnosis and disease monitoring. This study examined associations between automatically extracted speech/language features and NPS severity. METHODS: A total of N = 37 subjective cognitive decline and N = 20 mild cognitive impairment participants from the BioBank Alzheimer Centre Limburg study were recorded performing a low-constraint free-speech task. NPSs were assessed using the Geriatric Depression Scale and the Neuropsychiatric Inventory. Acoustic and linguistic features were automatically extracted. Correlation analysis was performed (adjusted for age, sex, and Mini-Mental State Examination) between the features and clinical scales. RESULTS: Features correlated significantly with NPSs. Indicatively, depression correlated with local jitter (r = 0.38, p < 0.001) and agitation with the sum of pause duration (r = 0.32, p < 0.027). DISCUSSION: Speech analysis offers a promising tool for evaluating NPSs in NCDs. Highlights: We found links between neuropsychiatric symptom (NPS) severity and speech markers in memory clinic patients. Temporal markers positively correlated with the presence and severity of agitation. Depression was positively correlated with voice instabilities. Anxiety was negatively associated with metrics of lexical diversity. Speech analysis provides an objective tool to assess NPSs in subjective cognitive decline and mild cognitive impairment patients.