Now You Hear Me, Now You Don’t

You hear footsteps coming down the hallway and voices chanting, “We’re coming.” A shadowy figure suddenly appears in the corner of your room, while a girl dressed up as a rat looms over your bed. Although these scenarios sound like they come from horror movies, they actually are real examples of hypnagogic hallucinations, which occur during the onset of sleep. These encounters made Yale psychiatrist and neuroscientist Dr. Albert Powers and neuroscientist Dr. Philip Corlett fascinated with auditory hallucinations. Their research, featured in Science, suggests that people who hear voices are more likely to experience induced hallucinations in a lab.

It may seem concerning that both Powers and Corlett have experienced hallucinations while falling asleep, but these hallucinations are usually symptomatic of a neurological condition, not a psychiatric illness. However, people without a psychiatric condition can hear voices too. The two scientists wanted to figure out what produces auditory hallucinations and why some voice-hearing experiences are benign and others require medical attention.

To explore these questions, they sought out four groups of test subjects: both psychotic and nonpsychotic voice-hearers and non-voice-hearers. After identifying potential subjects, the researchers separated the psychotic and nonpsychotic people using a questionnaire developed by forensic psychologists to distinguish between people who were actually experiencing hallucinations and those who only claimed to do so.

After screening their subjects for hallucinations, Powers and Corlett induced auditory hallucinations in their subjects to identify whether psychotic people were more likely than non-psychotic people to hear conditioned sounds. Using a technique originally developed at Yale during the 1890s, the subjects were stimulated with a checkerboard image and a one-second long sound simultaneously and repeatedly, while getting their brains imaged by MRIs. This conditioned the subjects to associate the image with the tone. As the scientists changed the intensity of tone, sometimes turning it off, the subjects pressed a button when they thought they heard the tone, changing the length of time they pressed the button to show their level of confidence.

Many reported hearing a tone when only the checkerboard image appeared but no tone played. This inconsistency occurred more often with the two voice-hearing groups, the people with schizophrenia and self-identified clairaudient psychics. Both groups were almost five times more likely to report that they heard a nonexistent tone than the non-voice-hearing groups. Furthermore, the two non-voice-hearing groups were 28% more confident that they had heard the tone when no tone played. These results support a possible explanation for hallucinations. “The brain makes models for what the outside world is like,” said Corlett, noting that these models sometimes don’t always match reality. This study suggests that people hallucinate when their expectations overweigh what their senses tell them.

Powers and Corlett further understood auditory hallucinations through analyzing the MRI scans collected: the parts of the brain that were responsive to the tone were active when people reported conditioned hallucinations, producing MRI scans of brains that looked like those of people actually hearing the tone. The images also revealed that both hallucinating and non-hallucinating people with psychosis exhibit abnormal brain activity in regions that monitor internal representations of reality. These results contribute to the idea that hallucinations stem from internal representations overruling actual sensory data.

This study was able to distinguish not only between people who hallucinate and people who don’t, but also between psychotic and non-psychotic people. “The sooner you catch psychosis and the sooner you intervene, the better the general outcomes are,” said Powers. According to Powers, most people with the symptoms associated with increased chances of psychosis don’t even develop psychosis. The question then is, who should receive treatment? This new research may help to answer that key question, by providing the basis for tests to diagnose patients who require psychiatric treatment early.

“There is no all-size-fits-all anti-psychotic, so there should be different treatments for different people,” Corlett said. Although this sort of precision medicine has not existed in psychiatry so far, Corlett hopes that this study, along with future research, will lead to more personalized psychiatric treatments, which he believes would be more effective in helping people suffering with mental health issues.

Featured photography by Yasmin Alamdeen