By Emma Young
Hallucinations are a common symptom of schizophrenia and related disorders, but mentally well people experience them, too. In fact, work suggests that 6-7% of the general population hear voices that don’t exist. However, exactly what predisposes well people to experience them has not been clear. Now a major new study of 1,394 people native to 46 different countries, led by Peter Moseley at Northumbria University, provides support for two hypotheses from earlier, smaller studies — namely, that a history of childhood trauma and a propensity to hear non-existent speech among background noise are both associated with experiencing hallucinations — but does not support three others.
“In terms of reproducibility, these results may be a cause for concern in hallucinations research (and cognitive and clinical psychology more broadly),” writes the team in their paper in Psychological Science. In firming up a few ideas, the work does, though, help to clarify what aspects of cognition as well as past experience are — and are not — linked to being more prone to hallucinations.
The participants came into one of 11 data collection labs (in the UK, France, the Netherlands, the Czech Republic, Canada, Norway and Australia) or participated online. They completed two scales that measured hallucinatory experiences, such as hearing, seeing or smelling something when there was nothing to explain those perceptions. They also reported on incidences of childhood trauma (being regularly criticised by a parent, for example). And they completed a series of tasks that tapped into cognitive processes identified in earlier work as being linked to hallucinations in the general population, as well as in patient groups.
The team found that experience of childhood trauma and performance on just the “auditory signal detection task” was linked to hallucinations. That task (which was completed by only the 594 participants who visited the labs) measured the participants’ ability to tell whether or not brief clips of speech had been played during longer bursts of noise. Those who scored higher on the hallucinatory scales had a higher false alarm rate — they were more likely to report hearing a voice when none was present. This data supports the previously proposed idea that hallucinations are linked to the brain over-relying on expectations of what will be perceived, vs actual sensory input. A higher false alarm rate on this kind of task is also seen in patients with schizophrenia, as is a history of adverse experiences in childhood.
However, in contrast to what has been found for patient groups and smaller, earlier studies of the general population, there was no association between hallucinatory experiences and results on tests of “dichotic listening” (which assessed the degree to which language processing was lateralised), “source memory” (remembering whether they’d actually heard or only imagined hearing various words) or verbal working memory. So, it seems that for well people who hallucinate, a failure to distinguish between what they have imagined vs what they have actually sensed, for example, doesn’t seem to be a cause. This adds to recent doubts about the idea of a “continuum model” for voice-hearing, which theorises that patients are simply further along the same voice-hearing spectrum as people in the general population who hear voices.
However, given the previously small sample sizes, lack of standardisation of studies on patients, and sparsity of direct replications, it’s hard to be sure whether hallucinations in well vs mentally unwell people have fundamentally different causes. “Further preregistered studies with large samples in these groups are needed,” the team writes. That work would not only help to further address the reproducibility issues in this field of research, but hopefully also clarify the mechanisms underlying hallucinations in general.