An intricate relationship
The exploration of conceptual mental processes throughout history has provided a valuable framework for understanding the intricate relationship between the mind, body, and the environment that surrounds us.
From the early philosophers of ancient Greece, who pondered the nature of consciousness and the physical embodiment, to modern theories on the extended and embodied mind, our understanding of ourselves and the world has been continually shaped and refined.
In the late 1990s Andy Clark and David Chalmers introduced the theory of extended mind, which proposes that the mind is not confined to the boundaries of the skull but extends beyond the body through interactions with the external environment. This perspective, along with the emerging theory of predictive mind in neuroscience, suggests that the human brain constantly generates predictions about the surrounding world to guide our behavior and perceptions.
The predictive brain functions not as a passive inference engine confined within the head, but as an active engagement engine that continuously interacts with the environment, providing a continuous grasp on task-relevant opportunities. Deep cognitive engines sculpt and maintain this grip through dynamic prediction, enabling us to navigate a world tailored to our needs. Equipped with estimations of the precision of our predictions, individuals proficient in predictive processing (PP) excel at discerning salient causes amidst sensory uncertainty.
Perception and action emerge as intertwined aspects of a unified computational process, rooted in multilevel prediction-error minimizing routines. Within this complex circular flow, actions select sensory stimulations that both test and respond to bodily and environmental cues. Knowledge-driven systems leverage structured probabilistic know-how encoded in multilevel generative models, optimizing strategy selection.
The integration between PP and embodied cognitive science is fully realized through perception-action loops and the incorporation of external resources within ongoing circular causal flows. As we explore the world, interoceptive and exteroceptive information converge, offering insights into our experiences, emotions, and affect.
The sensorimotor flow driven by PP provides a new understanding of perception, where experience, expectation, uncertainty, and action intertwine. By adapting to prediction errors and learning generative models, PP approaches build upon previous work in artificial neural networks and cognitive science, striking a balance between simplicity and complexity while adapting to different tasks and contexts.
The impact of higher-level cognitive processes on sensory perception originates from the development and utilization of predictive models that aim to anticipate incoming sensory information.
A classic illustration of this phenomenon can be found in the Ranshaw Cow Card image. Initially, most individuals may perceive only a basic pattern of light and shadow in the image. However, once deciphered, this understanding fundamentally alters our perception of the image, influencing how we perceive it for the rest of our lives.
In such instances, our pre-existing beliefs, implemented through the brain's predictive models, significantly shape our perception. These beliefs often operate at a subconscious level and are not necessarily accessible to conscious awareness. Instead, they manifest as a range of sub-personal states, primarily expressed in probabilistic rather than definitive terms.
It is time to recognize that our mind is part of a larger system that includes the environment and the tools we use.