Is abnormal inner speech the basis of auditory verbal hallucinations, and what evidence supports or challenges this account?

Why would your own inner speech sound like someone else?

The core idea is simple. When you silently talk to yourself, your brain also generates a prediction of what that inner speech will feel like, so it can tag the experience as self-produced. Researchers call this predictive signal 'corollary discharge,' and the tagging process 'self-monitoring' [2]. If the prediction is weak or mistimed, the inner speech may arrive without a 'this is mine' label, and the brain could treat it as coming from outside. That is the misattributed inner speech account: a voice is your own inner speech that failed to be recognized as yours [1][2].

The account is attractive because it explains several features of the experience at once: the voices are usually in the person's own language, they often comment on or argue with the person, and they feel external despite having no external source. A narrative review of phenomenological, cognitive, and neurobiological work frames the question as how far inner speech abnormalities actually go toward explaining hallucinations, and where the model needs to be integrated with other levels of explanation [1].

What the brain recordings actually show

The most direct evidence comes from electrical recordings of the brain during inner speech tasks. In one line of work, the suppressive mechanism that normally dampens self-generated inner speech was found to be dysfunctional in schizophrenia, which fits the idea that inner speech is not being properly tagged as self-produced [2]. A separate study measured a brain wave called the contingent negative variation, which builds up just before a person produces a movement or an inner syllable. Healthy controls showed a distinct preparatory pattern when they imagined a syllable compared with when they simply listened; patients with schizophrenia showed an abnormal pattern in both conditions, and the specific pattern differed between patients who currently heard voices and those who did not, significantly predicting who had hallucinations [3].

Older imaging work points the same way. When researchers compared hallucinating patients, non-hallucinating patients, and controls on inner speech tasks, hallucinations were associated with a failure to activate the brain areas normally concerned with monitoring inner speech [5]. Taken together, these studies are consistent with a monitoring failure during inner speech, but they measure different things — suppression, motor preparation, and regional activation — so they support the same broad story rather than proving one single mechanism.

The link is stronger in some patients than others

The relationship between inner speech and voices is not uniform. In a study of 189 chronic schizophrenia patients, the experience of hearing other people's voices within one's own inner speech was linked to more hallucinations — but only in people with low or moderate cognitive performance. In patients with high cognitive function, that association was not significant [4]. This suggests the inner speech pathway may matter most when other cognitive resources are already strained, and it is a reminder that the account describes a contributing route rather than a universal cause.

That moderation finding also has a practical edge: if cognitive performance shapes the link, then interventions aimed at improving cognition might also reduce hallucinations, though the authors describe this as preliminary [4]. The result comes from a single cross-sectional sample in Lebanon, so it identifies a pattern worth testing rather than a settled treatment target.

What the inner speech account still has to explain

Two puzzles keep the model from being the whole answer. First, inner speech in schizophrenia is not uniformly negative, yet the voices patients hear are more often negatively toned. One ongoing study is testing whether emotionally charged inner speech is harder for a compromised monitoring system to suppress, which could produce a 'breakthrough' into hallucination [2]. That is a hypothesis under test, not an established result. Second, the model needs to connect monitoring failures to the full experience of voices — their content, their emotional charge, and why they persist — which is why the narrative review argues for integrating inner speech with predictive processing and other levels of explanation rather than treating it as a standalone cause [1].

So the honest answer is that abnormal inner speech is a well-supported contributor with a plausible mechanism, backed by converging physiological evidence, but it is not yet a complete basis for auditory verbal hallucinations. The evidence is strongest for a monitoring or prediction failure during inner speech in schizophrenia-spectrum patients who hear voices [2][3][5], and weakest for the claim that this alone explains who develops voices and what those voices are like [1][4].

Sources used in this answer

Related research questions