Understanding auditory verbal hallucinations: a systematic review of current evidence

Abstract
Objective: Auditory Verbal Hallucinations (AVHs) are core features of psychotic illness, and remain significant in predicting poor outcome and risk. There has been a wide range of approaches to understanding these experiences.
Method: A systematic literature review summarising different methods of investigation and their results; phenomenology, descriptive psychopathology, psychological, cognitive neurobiology and neuroimaging.
Results: 764 papers and texts were screened and 113 reviewed. Phenomenological studies are comparably few in number and psychopathology remains based on concepts defined in the early 20th century. Psychological models focus on voice content and emotional reaction, and suggest a continuum of AVHs from normal experience. Neuropsychological models include AVHs as misattribution of inner speech, whilst functional neuroimaging studies focus on the spontaneous activity and connectivity of auditory networks.
Conclusions: There has been a large growth in research on AVHs in recent decades dominated by neurobiological and neuroimaging studies. Future research should include focus on phenomenological aspects and AVHs change over the course of developing illness. Integration between branches of enquiry is needed and the risk is that without this, models are proposed and investigated that bear scant relevance to the symptom itself.
Summations:
•Psychological models stress the importance of content and emotion
•Dimensional models propose AVHs as a continuum of normal experience
•Cognitive neurobiological models of AVHs include misattribution of inner speech, adherent memory, semantic processing errors, and abnormal connectivity
Considerations:
•Modern investigations of what constitutes the core features of AVHs are under-represented.
•Many models do not account for the rich complexity of AVH experience or how these may change over the development of a psychotic illness.
• Studies examining voice hearing in non-clinical samples have not been widely replicated and it remains unclear as to how relevant neurobiological models are in non-clinical and non-psychotic populations.