Whether they have been diagnosed with psychosis or not, people who hear voices are more sensitive than others to a century old experiment designed to induce hallucinations.
The hope is that studying people's ability to learn that these hallucinations aren't real could help identify those in need of psychiatric treatment.
In a study published in the journal Science, a team of researchers from Yale University set out to identify factors that contribute to auditory hallucinations.
"We wanted to test the hypothesis that hallucinations may be driven by prior beliefs, that hallucinations may arise from an imbalance between people's expectations about the environment and about the information they get from their senses," senior author Philip Corlett, an assistant professor of psychiatry at Yale, told IBTimes UK.
The scientists describe in their paper how they have used a form of Pavlovian conditioning, a method developed by the Russian physiologist Petrovich Pavlov in the late 19th century.
They conducted an experiment involving four different groups of participants - people who heard voices (both psychotic and non-psychotic) and people who didn't (bpsychotic and non-psychotic).
All were repeatedly presented with a visual stimulus associated with tone, while undergoing brain scans. They were told to detect the tone, which was difficult to hear at times. The researchers indeed varied the intensity of the tone and sometimes did not play it at all.
They found that after a while, people in all groups reported hearing a tone when the stimulus was presented, even though no tone was actually played.
The scientists noticed however that this effect was much more pronounced in the two groups that were prone to hearing voices. They were much more likely to feel strongly that they heard a tone when none was presented.
This suggests that people who are more prone to auditory hallucinations are also more sensitive to the experiment. They learnt more quickly to associate the tone to the stimulus, and they were overly influenced by expectations and prior associations.
"It's an interesting study because usually we differentiate people who have hallucinations from people who don't simply by asking them directly about their experiences or by observing them. In this case, the researchers were able to identify hallucinators experimentally - that is, they were able to directly elicit hallucinations with their task", Dr. Ian Kelleher a psychiatrist with the Royal College of Surgeons in Ireland, who was not involved with the study, told IBTimes UK.
"Compared to other animals, humans excel at learning, and here people with hallucinations appear to be learning very quickly - too quickly, in fact, so that they're actually jumping to conclusions by always pairing the sound with seeing the stimulus. We know from before that people with psychosis can jump quickly to conclusions too in other tasks and this study is showing this in a unique way. "
In need of treatment
Using computational models, the researchers also showed that participants with a psychotic illness had more difficulty accepting that they had not really heard a tone. Brain scans also revealed that they exhibited altered activity in brain regions often implicated in psychosis.
These findings constitute behavioural and neuroimaging markers that could help identify people who may go on to need psychotic treatment.
"In certain early stages of the psychotic illness, it can be ambiguous if someone will need treatment or not. Identifying those who will is very important, especially considering that early intervention is key to help people with psychosis," study lead author Al Powers, a clinical instructor in psychiatry said.