Hallucinations occur in 6-15% of the general population, but their detailed characteristics are poorly understood. In a large online survey of 10,448 participants aged 14-88, auditory hallucinations were most common in the past month (29.5%), followed by visual (21.5%), tactile (19.9%), and olfactory (17.3%); nearly half (47.6%) experienced hallucinations in two or more senses. Many participants rated their hallucinations as severe due to negative content (16.0-31.6%), distress (10.5-16.8%), or dysfunction (12.7-17.3%). Decreased insight occurred in 10.2-11.4%, hypnagogia in 9.0-10.6%, and bereavement hallucinations in 2.8%. Delusions were rare (7.0%) but significantly associated with recent hallucinations. The findings indicate a wide variety of hallucination phenomenology in the general population and support a phenomenological continuum.
Hallucinations, which occur in many psychiatric disorders, may be better understood through the lens of developmental psychology. Their clinical significance depends on when they appear in a person's life. Key cognitive-developmental processes—such as engaging with imaginary entities, adverse events, executive functioning, social cognition, and language development—shape how hallucinations arise across different sensory modalities. Atypical developmental trajectories, as seen in certain conditions, also influence hallucination prevalence and phenomenology. Integrating developmental and psychiatric perspectives could yield mutual benefits for future research.
A scoping review of 10 studies found that artificial intelligence methods, including machine learning and digital phenotyping via smartphones and wearables, show promise for detecting relapse in psychotic disorders but have significant limitations. The sensitivity of AI models ranged from 0.25 to 0.77 and specificity from 0.06 to 0.88, with area under the curve between 0.63 and 0.78. Models were heterogeneous and most findings were not replicated. The review concludes that while personalized approaches with individual-level modeling are promising, larger studies and methods such as large language models are needed before AI can be used in real-world clinical practice.