A time-series network approach to auditory verbal hallucinations: Examining dynamic interactions using experience sampling methodology

Background: Identifying variables that influence daily-life fluctuations in auditory verbal hallucinations (AVHs) provides insight into potential mechanisms and targets for intervention. Network analysis, that uses time-series data collected by Experience Sampling Method (ESM), could be used to examine relations between multiple variables over time. Methods: 95 daily voice-hearing individuals filled in a short questionnaire ten times a day for six consecutive days at pseudo-random moments. Using multilevel vector auto-regression, relations between voice-hearing and negative affect, positive affect, uncontrollable thoughts, dissociation, and paranoia were analysed in three types of networks: between-subjects (between persons, undirected), contemporaneous (within persons, undirected), and temporal (within persons, directed) networks. Strength centrality was measured to identify the most interconnected variables in the models. Results: Voice-hearing co-occurred with all variables, while on a 6-day period voice-hearing was only related to uncontrollable thoughts. Voice-hearing was not predicted by any of the factors, but it did predict uncontrollable thoughts and paranoia. All variables showed large autoregressions, i.e. mainly predicted themselves in this severe voice-hearing sample. Uncontrollable thoughts was the most interconnected factor, though relatively uninfluential. Discussion: Severe voice-hearing might be mainly related to mental state factors on the short-term. Once activated, voice-hearing appears to maintain itself. It is important to assess possible reactivity of AVH to triggers at the start of therapy; if reactive, therapy should focus on the triggering factor. If not reactive, Cognitive Behavioural interventions could be used first to reduce the negative effects of the voices. Limitations are discussed. (C) 2019 Elsevier B.V. All rights reserved.