In traversing the labyrinth of the human mind, scientists stumble upon fascinating theories and paradigms that act as illuminating lanterns. One such theory that has grabbed attention and incited immense curiosity is predictive coding – the brain’s unique mechanism to make sense of the world.

In simple terms, predictive coding propounds that our brains are prediction machines which often decode sensory information as well-anticipated events. This article aims to delve deep into the intricate workings of predictive coding and analyse its impact on our understanding of the human brain and cognitive processes.

As a conceptual model, predictive coding originated in the late 1980s(1), with the word ‘coding’ borrowed from the realms of telecom and computer science. It became popular due to Geoffrey Hinton and Karl Friston’s foundational work that meticulously investigated the neural basis and implications of this theory(2).

Predictive coding’s governing axioms are rooted in the hierarchical structure of the brain. In this structure, each level predicts the level below it, and the prediction errors also move up the hierarchy. This hierarchical prediction manifests in all aspects of our sensory experiences, from visual to auditory to tactile.

Predictive coding’s influence is palpable in neuroscience, cognitive science, and artificial intelligence, where it molds our understanding of perception, attention, and learning. This theory also bolsters Bayesian brain predictions, postulating that the brain is adept at performing potentially complex Bayesian inferences to make sense of the world around us(3).

The enigmatic nature of predictive coding is best encapsulated in the words of philosopher Jakob Hohwy, who argued, “We don’t just passively receive information from the world. We actively predict and explain it, all under the brain’s influence”(4).

Predictive coding has ignited controversies in the academic world. Critics feel predictive coding oversimplifies how our brains work, suggesting that not all brain processes are inherently predictive. Others voice concerns that the Bayesian brain predictions are mathematically complex and question how these concepts could be implemented in the neural circuits of the brain(5).

Peer into the world of AI and machine learning, and predictive coding’s influence becomes starkly visible. Advances in deep learning reflect hierarchical predictive coding principles where neural networks employ layers of processing units to predict input from the previous layer(6). This attempt to mime the human brain, however, is still in its infancy.

Delving into implications, this theory might hold the potential to revolutionise how we diagnose and treat mental ailments. Aberrations in predictive processing could be the underlying cause for conditions like autism, schizophrenia, and anxiety disorders. This perspective could open up new dimensions towards effective treatment techniques(7).

What’s next for predictive coding is intriguing. Propelled by advancements in neuroscience and artificial intelligence, we can anticipate further research to decode the brain’s uncanny prediction abilities. As scientists embark on this journey, the ultimate destination is yet to be reached.

If the findings of predictive coding resonate with the rise of AI and machine learning, could it be that our brains, the palpable epitomes of biological intelligence, are actually a form of ‘biological machine learning’? To quote a pioneer in machine learning, Yoshua Bengio, “These ideas are changing the way we understand intelligence”(8). That, indeed, opens a door for thought to us all.

References and Further Reading

1. Mumford, D. (1992). On the computational architecture of the neocortex. Biological Cybernetics, 66(3), 241–251.
2. Friston, K. (2009). The free-energy principle: a rough guide to the brain?. Trends in cognitive sciences, 13(7), 293-301.
3. Rao, R. P., & Ballard, D. H. (1999). Predictive coding in the visual cortex: a functional interpretation of some extra-classical receptive-field effects. Nature Neuroscience, 2(1), 79-87.
4. Hohwy, J. (2013). The Predictive Mind. Oxford University Press.
5. Bowers, J. S., & Davis, C. J. (2012). Bayesian just-so stories in psychology and neuroscience. Psychological Bulletin, 138(3), 389-414.
6. Schmidhuber, J. (2015). [title of the document] Deep learning in neural networks: An overview. Neural Networks, 61, 85-117.
7. Clark, A. (2013). Whatever next? Predictive brains, situated agents, and the future of cognitive science. Behavioral And Brain Sciences, 36(3), 181-204.
8. Bengio, Y. (2017). Deep Learning of Representations for Unsupervised and Transfer Learning. Journal of Machine Learning Research, 17, 1-5.