Perceptual inference.

Area of Science:

• Cognitive Neuroscience
• Computational Neuroscience
• Psychology

Background:

• Perception involves inferring sensory stimuli using internal neural representations shaped by experience.
• Bayesian inference and decision theory offer frameworks for modeling cognition, utilizing error sensing and generative models.
• Perception is viewed as information evaluation, distinct from sensation, relying on stored memories and guided by sensory feedback.

Purpose of the Study:

• To explore the concept of perceptual inference as a fundamental cognitive process.
• To elucidate the role of internal models and memory in shaping perception.
• To examine the neural mechanisms underlying perceptual inference, particularly in the cerebral cortex.

Main Methods:

• Conceptual analysis integrating Bayesian inference, decision theory, and neuroscience.
• Review of evidence from phenomena like cortical co-localization of perception and memory.
• Examination of neuronal response invariance and dissociation of perception from sensation.

Main Results:

• Perception is characterized as an active process of evaluating sensory information against internal models.
• Stored representations (memories) function as internal models for predicting and associating sensory stimuli.
• The cerebral cortex operates akin to a servomechanism, employing sensory feedback to refine internal models.

Conclusions:

• Perceptual inference is a key mechanism for constructing our understanding of the external world.
• The brain continuously updates its internal models based on sensory feedback and past experiences.
• Understanding perceptual inference provides insights into the relationship between memory, prediction, and conscious experience.