Predictive Processing Explains Content of Charles Bonnet Hallucinations

Visual hallucinations experienced by individuals with Charles Bonnet Syndrome (CBS) result from a combination of sensory pathway damage and the brain's predictive mechanisms shaping perception. CBS is characterized by complex visual hallucinations involving fully formed objects unrelated to psychiatric disorders, often accompanied by insight into their unreality.

These hallucinations typically occur in people whose vision loss stems from damage to sensory inputs reaching the brain. For example, the author experienced recurring hallucinations of people, bicycles, and keyboards during specific activities such as learning to use a white cane or working on a laptop. These repetitive hallucinations raise questions about why certain images appear repeatedly and why they eventually cease.

How Visual Perception Operates

In sighted individuals, visual perception begins when light activates receptors in the eyes, sending neural signals to the cerebral cortex's visual areas. The brain's visual system is hierarchically organized, with early visual cortex regions processing basic sensory features and higher-level areas interpreting more abstract and complex elements.

Research has demonstrated that the content of CBS hallucinations corresponds to the perceptual features processed by specific higher-level cortical areas. For instance, hallucinations of faces activate regions specialized for face perception. Visual perception involves an integrated network with bidirectional information flow, allowing interactions between different hierarchical levels.

This bidirectional flow supports two main perceptual processes: bottom-up processing, which analyzes sensory input into components, and top-down processing, which applies prior knowledge to resolve ambiguities. Perception thus involves interpretation rather than mere reproduction of sensory data. For example, prior knowledge about poisonous snakes may lead to misinterpreting a fallen branch as a snake, explained by predictive processing theories.

Predictive Processing and Hallucination Formation

Predictive processing theory posits that conscious visual perception is the brain's best inference about sensory input based on past experiences. CBS hallucinations can be understood by combining this theory with deafferentation theory, which attributes hallucinations to loss of sensory input to brain perceptual areas.

Deafferentation may cause the visual system to become hyperexcitable, lowering activation thresholds and generating spontaneous activity without external stimuli. This instability could produce hallucinations; however, it does not fully explain why specific hallucinations recur. Predictive processing suggests that learned expectations influence the form of hallucinations by shaping internally generated neural activity.

Role of Learned Expectations in Hallucination Content

Activities that are well learned, such as typing, create contexts where the brain expects certain sensory inputs. For example, preparing to type may lead the brain to predict the presence of a keyboard based on prior visual, tactile, auditory, and proprioceptive experiences. When actual sensory evidence closely matches these predictions, the brain infers a keyboard's presence, shaping spontaneous neural activity into a hallucinated image.

This process aligns with the idea that perception involves constructing what would be necessary for sensations to make sense. Over time, neural plasticity allows the brain to reorganize and reduce the instability that facilitates hallucinations, leading to their eventual disappearance in most individuals.

Scientific References Supporting CBS Hallucination Theories

The explanation of CBS hallucinations draws on multiple studies and reviews. Altieri and Battaglini (2026) discuss neural mechanisms beyond hyperexcitability. Clark (2024) explores how minds predict and shape reality. Ffytche et al. (1998) link hallucination content to specific cortical areas. Mars et al. (2025) provide an overview of CBS, while Marschall et al. (2020) examine deafferentation as a cause of hallucinations. Pang (2016) reviews hallucinations in visually impaired individuals, and Peelen et al. (2024) analyze predictive processing of scenes and objects. Powers et al. (2016) describe hallucinations as top-down effects on perception.