Despite the desire to delve deeper into hallucinations of all types, methodological obstacles have frustrated development of more rigorous quantitative experimental techniques, thereby hampering research progress. Here, we discuss these obstacles and, with reference to visual phenomena, argue that experimentally induced phenomena (e.g. hallucinations induced by flickering light and classical conditioning) can bring hallucinations within reach of more objective behavioural and neural measurement. Expanding the scope of hallucination research raises questions about which phenomena qualify as hallucinations, and how to identify phenomena suitable for use as laboratory models of hallucination. Due to the ambiguity inherent in current hallucination definitions, we suggest that the utility of phenomena for use as laboratory hallucination models should be represented on a continuous spectrum, where suitability varies with the degree to which external sensory information constrains conscious experience. We suggest that existing strategies that group pathological hallucinations into meaningful subtypes based on hallucination characteristics (including phenomenology, disorder and neural activity) can guide extrapolation from hallucination models to other hallucinatory phenomena. Using a spectrum of phenomena to guide scientific hallucination research should help unite the historically separate fields of psychophysics, cognitive neuroscience and clinical research to better understand and treat hallucinations, and inform models of consciousness.
This article is part of the theme issue ‘Offline perception: voluntary and spontaneous perceptual experiences without matching external stimulation’.