How robots make people hear voices

© iStock | Benjavisa

Researchers have developed a method for triggering auditory hallucinations in healthy people. They are investigating the causes of a phenomenon that can be a huge burden for people with psychiatric illnesses.

Studies indicate that five to ten per cent of all people sometimes hear voices – those of dead relatives or higher beings, for example – without having any kind of illness. In other words, the phenomenon is not restricted to people with psychiatric disorders, as is generally assumed. However, hallucinations do frequently accompany such illnesses. “More than seventy per cent of people with schizophrenia hear voices,” SNSF-funded neuroscientist Pavo Orepic tells us. And the bad thing about it: what the voices say is generally very negative. “Sometimes they talk to people daily and make them believe awful things about themselves, for example that they’re worthless and should self-harm.”

As a member of Olaf Blanke’s team at EPFL, Orepic has developed a procedure that can artificially make people hear voices. As a result, the researchers can now investigate how auditory hallucinations occur in healthy people and identify possible treatment approaches. While the psychopharmaceuticals used up to now alleviate the symptoms to a certain extent, they have a non-specific effect on many systems within the body and cause severe side effects. Elucidating the mechanisms behind auditory hallucinations could lead to the development of specifically acting treatments.

“We still have no idea what happens in the brain when people have auditory hallucinations,” Orepic tells us. Some studies indicate that they can occur when sensory impressions do not match the brain’s expectations – for example if you put a bakery product that you expect to be sweet in your mouth only to find that it is savoury. Other investigations found an alternative explanation – that hallucinations occur when the brain has been conditioned by previous impressions and interprets sensory perceptions incorrectly as a result.

An experiment to deliberately confuse the brain

Now Orepic has designed an experiment that triggers these two mechanisms simultaneously. The people taking part in the experiment are blindfolded, then asked to press a lever in front of them. As they do so, a robot that they cannot see prods them in the back. As a result, the sensory impression they receive does not align with what their brain expects, as described in the first theory. Earlier studies have shown that over time, this process gives the test subjects the illusion that they are touching their own back.

Once they have got used to this experience, the prod is delayed very slightly. “Now the brain explains that delayed sensory feedback as someone else being present and touching them in the back,” Orepic says. According to the second theory, incorrectly perceiving a presence in this way can lead to hallucinations. To determine whether the procedure encourages auditory hallucinations, the researchers then played the test subjects noises into which they had mixed either very soft voices – sometimes the subjects’ own, sometimes someone else's – or no voice at all. They found that after the “delayed prod” procedure, the subjects were more likely to hear other people’s voices among the noises compared to control conditions even if no voice had actually been mixed in.

A mechanism inherent in all of us

“The result shows that the two theories on how hallucinations occur are not mutually exclusive but might mesh with each other instead,” Orepic says. It is also a major step forward, he says, that the method can be used to investigate the phenomenon of hearing voices in healthy people under controlled conditions. This is because it is difficult to interpret the results of experiments involving people who hear voices as a result of illness because of interference from medicines or other effects. “Our study confirms that the mechanisms behind the hallucinations are actually in everyone’s brain,” Orepic says. “But for some reason, some people are more susceptible to them than others.”

Orepic believes that the threshold between harmless and pathological hallucinations is fluid anyway: one criterion could be whether hearing voices impacts people’s lives negatively, making them self-harm, for example, or preventing them from leading autonomous lives. Conversely, voices can be harmless or even benign – for example if someone hears the voice of their dead grandmother who offers good advice.

Orepic hopes that his research will ultimately help remove the stigma from people who hear voices. “That can only happen if we find out more about the causes of hallucinations and raise public awareness of them”.