Generated by Codex with GPT 5.6 Sol XHigh

When an Inner Voice Sounds External

Most people can imagine saying a word without mistaking the thought for a sound in the room. The brain appears to preserve that distinction partly by predicting the sensory consequences of self-generated speech. When a person speaks or prepares to speak, motor regions involved in producing the sound suppress activity in the auditory cortex. That dampening helps label the resulting sound as one’s own rather than something arriving from outside.

In “Selective Sound,” Hannah Seo describes evidence that this predictive system also operates during silent inner speech—and that it behaves differently in some people with schizophrenia who experience auditory hallucinations. Such hallucinations are common across schizophrenia-spectrum disorders, but they are neither universal nor a simple measure of how severe a person’s condition is. The study investigates one possible neural mechanism behind hearing voices, not a complete explanation of schizophrenia.

Listening to Imagined Syllables

Inner speech is difficult to study because no observer can directly verify what another person is imagining. Cognitive neuroscientist Thomas Whitford and his colleagues approached the problem by combining a tightly controlled task with electroencephalography, or EEG, which records electrical activity in the brain.

The study included adults without schizophrenia and people on the schizophrenia spectrum, some of whom were currently hearing voices and some of whom were not. Participants were prompted to imagine saying either “bah” or “bee” without moving their mouths. At the same time, headphones played a syllable that either matched or differed from the imagined one. In control trials, participants simply listened without imagining speech.

Among adults without schizophrenia, imagining and hearing a sound at the same time reduced the auditory cortex’s response compared with listening alone. The dampening was strongest when the heard and imagined syllables matched. This is what a predictive system should do: the closer an incoming sound is to the brain’s internally generated expectation, the more confidently the brain can mark it as self-produced and turn down its response.

Participants with schizophrenia who were experiencing auditory hallucinations showed the reverse pattern. A matching imagined and heard syllable produced a stronger auditory response rather than greater suppression. Their inner speech appeared to amplify the very signal that the comparison group’s brains dampened. If the same disruption occurs outside the laboratory, a privately generated thought might carry less of the neural signature that normally identifies it as one’s own.

The group with schizophrenia who were not currently hearing voices fell between these patterns. Mismatched sounds produced strong dampening, but matched sounds did not. Whitford suggests this may reflect a latent susceptibility to hallucinations, although the result does not establish that interpretation or show that hallucinations will develop.

A Clue, Not a Diagnostic Test

The findings support a long-standing idea: auditory hallucinations can arise when communication between speech-planning and auditory regions fails to distinguish inner from external sound. They also provide direct evidence that silent imagined speech can alter auditory-cortex activity. The article notes, however, that researchers do not yet know whether this pattern contributes to every kind of sound-based hallucination, much less to the many other symptoms and pathways encompassed by schizophrenia.

The experiment should therefore be read as a mechanistic clue rather than a clinical test. The article does not report the study’s sample sizes, and its laboratory task reduces inner speech to imagined syllables paired with actual sounds. Everyday voices are far more complex: they can have identities, emotional force, conversational structure and personal meaning. A neural response to “bah” or “bee” cannot capture that full experience.

Whitford hopes that a refined EEG measure might eventually help identify people at risk of psychotic symptoms early enough for preventive care. That possibility remains prospective. For now, the result’s value is more fundamental: it shows that the boundary between thought and perception is actively constructed. Ordinarily the brain marks its own predictions and quiets their sensory echo. When that marking system changes, an inner event may be experienced with some of the force of an external voice.