Hearing what isn’t said: The brain’s illusion of internal voices

Every time he closes his eyes, the image of his father appears, a stern face etched into childhood memory. And with it, a harsh, violent voice. No one is speaking. He’s alone in the room. But he hears it clearly, insults, accusations, phrases that seem to emerge not from the outside world, but directly from the mental image itself.

Experiences like this are not rare. They’re classic examples of verbal auditory hallucinations (VAHs), a phenomenon where the brain treats an internal representation as if it were an external perception. This confusion, between thoughts generated by the mind and real sensory input, represents one of the brain’s most intriguing failures in source recognition. Long seen as a hallmark of psychotic disorders, VAHs are now acknowledged to occur in far broader contexts. They appear in individuals with anxiety, PTSD, mood disorders, and even in people with no psychiatric diagnosis. This diversity raises new questions about their origin and underlying mechanisms.

For several years, researchers have sought to understand why the brain “hears” what was never said. How can an inner image, a memory, a face, a scene, turn into a sensory experience? Why do some people experience these hallucinations so frequently, and in such intrusive ways? The leading hypothesis suggests that these phenomena aren’t simply due to an auditory malfunction, but result from a more complex interplay between memory, emotion, and mental imagery. A recent systematic review sheds new light on this topic by clearly distinguishing the roles of emotionally charged versus neutral imagery in the emergence of auditory hallucinations.

Can we really hear what we imagine?

It sounds paradoxical, even absurd. And yet, the question leads us to a core insight: some mental images, especially those laden with emotion, might not only activate the brain’s perceptual circuits, but also produce auditory experiences in the absence of any real sound. That’s exactly what a 2024 study by Janssen and colleagues explored, in collaboration with the universities of Maastricht, Radboud, and Oxford.

Their findings are striking: emotionally neutral mental images, such as visualizing objects or neutral scenes, are not associated with verbal auditory hallucinations. In contrast, emotionally charged images, often intrusive and tied to traumatic memories or negative self-representations, play a central role. Around 70% of participants who experienced VAHs reported such imagery at the moment the hallucination occurred.

Conversely, positive mental imagery, when intentionally summoned, appears to have a calming effect. In some cases, it even seems to alleviate symptoms, particularly those linked to depression. These findings suggest that it’s not mental imagery per se that leads to hallucinations, but rather the emotional charge it carries, and above all, the way it imposes itself on consciousness, often involuntarily and intrusively.

Inside the brain: The neural roots of inner voices

Verbal auditory hallucinations aren’t caused by a single sensory glitch. Instead, they result from functional disruptions in interconnected brain systems that normally allow us to distinguish between thoughts, mental images, and real-world perception. These systems include networks involved in memory, language, emotion, and reality monitoring.

At the heart of this process is the superior temporal gyrus, a region of the auditory cortex that processes speech and sound. Under normal conditions, it responds to external auditory stimuli, voices, phonemes, environmental sounds. But during hallucinations, this region may become spontaneously active, even in the absence of external input. This endogenous activation reveals a profound fact: the brain doesn’t merely perceive passively; it can generate its own sensory content from within.

The problem arises when the brain’s reality-monitoring system, particularly the medial prefrontal cortex, fails to function correctly. This area plays a key role in cognitive self-monitoring, enabling us to determine whether a thought or experience is internally generated or comes from the outside. When this “source recognition” mechanism breaks down, auditory representations from memory or inner speech may be misattributed to external sources. In other words, the brain hears what it has produced itself, without realizing it.

This misattribution is exacerbated when other structures, such as the amygdala and hippocampus, amplify the emotional intensity of mental images. A memory linked to a traumatic experience, reactivated in a moment of solitude or distress, gains vividness. It becomes salient, invasive, and crosses the threshold into conscious perception.

At that point, it’s no longer just a thought. It manifests as a voice, with a specific tone, a spatial presence, and a distinct auditory texture. Deprived of cues to recognize its internal origin, the brain treats the mental image as an external sound, triggering the same neural pathways involved in actual hearing.

This phenomenon reflects a rupture in the boundary between perception and imagination. What is typically contained in the private space of inner dialogue, memories, anticipations, imagined scenarios, becomes, in these moments, an external event. The brain confuses imagined content with real sensory input. Hallucinated voices, then, do not come out of nowhere. They are crafted from internal material, emotionally charged mental images, which the perceptual system mistakenly treats as reality.

When emotions shape what we hear

These neurocognitive mechanisms reveal a principle well known in psychology: our emotional state shapes our perception. A person under acute stress, guilt, or loneliness may experience mental images that are more vivid, verbal, and invasive. If these emotionally charged representations aren’t recognized as such, they can cross a threshold where they are misinterpreted as external voices. Perception, in this case, is constructed internally, but presented in the brain as coming from outside.

Still, it is important to distinguish the hallucinations observed in this study from those found in chronic psychotic disorders like schizophrenia. In the transdiagnostic framework studied by Janssen and colleagues, the reported voices were typically contextual, tied to trauma, emotional distress, or psychological vulnerability. They were often intermittent, partially recognized as self-generated, and responsive to psychotherapeutic interventions.

Persistent psychotic hallucinations, on the other hand, tend to be rigid, frequent, and embedded in a delusional system, with a marked loss of insight into their origin. Therefore, rather than interpreting all hallucinations as signs of a total break with reality, some can be understood as the result of mental processes pushed to their limit, a momentary breakdown in the brain’s ability to separate the inner world from the outer one.

Healing the inner world behind the voices

This shift in understanding marks a turning point. Rather than viewing voices as inherently pathological or irrational, recent research situates them within an intelligible neuropsychological framework. In non-psychotic contexts especially, this opens the door to new therapeutic strategies, ones that focus not on silencing the voices, but on exploring their emotional, cognitive, and mnemonic roots.

Instead of suppressing the symptom, treatment can target the internal representations behind it: traumatic memories, emotional imagery, distorted beliefs, and misattributions of source. By reducing the vividness of these images, altering their emotional tone, or helping the person recognize them as self-generated, the intrusive quality and persuasive power of the voices may diminish.

This is how research can lead to more refined therapies, tailored to the individual’s internal dynamics. Hallucinations are no longer treated as isolated anomalies, but as reflections of a broader imbalance in how mental representations are organized. The “mirror of voices,” then, becomes the real focus of care, a malleable space shaped by personal history, emotional vulnerability, and regulatory strategies.

By working with this underlying dynamic, therapeutic interventions can go beyond symptom management. They can foster deeper understanding and integration of the lived experience itself.

Reference

Janssen, H., van den Berg, K. C., Paulik, G., Newman-Taylor, K., Taylor, C. D. J., Steel, C., Keijsers, G. P. J., & Marcelis, M. C. (2024). Emotional and non-emotional mental imagery and auditory verbal hallucinations (hearing voices): A systematic review of imagery assessment tools. Clinical Psychology & Psychotherapy, 31(1), e2920.