Module 8

Applications & Theories of Mind

Bayesian-brain ideas increasingly appear in engineering (BCIs, world-model AI) and in theories of consciousness (GNW, IIT, AST, FEP, HOT). This final module surveys applications and positions the Bayesian brain relative to the principal consciousness theories.

1. BCI Decoding & World-Model AI

Modern brain-computer interfaces use Kalman filters and variational encoders to decode motor intent from M1 spiking or ECoG. Neuralink, BrainGate, and Blackrock systems all employ explicit Bayesian decoding pipelines. In AI, Dreamer (Hafner 2019) and its successors implement world-model RL that closely parallels active inference: sample-efficient policy learning via an internal generative model of environment dynamics.

2. Consciousness Theories

Several contemporary theories compete for the “explanandum” of consciousness:

  • Global Neuronal Workspace (GNW, Dehaene): conscious access = broadcast to a workspace.
  • Integrated Information Theory (IIT, Tononi): consciousness = Φ, a measure of cause-effect integration.
  • Attention Schema Theory (AST, Graziano): consciousness = the brain’s model of its own attention.
  • Free-Energy Principle (FEP, Friston): consciousness aspects = internal-state inferences meeting Markov-blanket conditions.
  • Higher-Order Theories (HOT, Lau): consciousness = meta-representation of first-order states.

Simulation: Theory Comparison

Python
script.py35 lines

Click Run to execute the Python code

Code will be executed with Python 3 on the server

3. Course Synthesis

The Bayesian brain is the most unified computational theory of perception, action, learning, and psychopathology available today. Modules 0–7 traced the framework from Bayes’ theorem through generative models, prediction-error precision, predictive coding, free energy, active inference, neural implementation, and computational psychiatry. Module 8 placed the framework in its broader computational and philosophical context. Whether the FEP is the unifying principle of life or an elegant formalism in search of predictive content is a question the next decade of neuroscience will help answer.

Key References

• Dehaene, S. (2014). Consciousness and the Brain. Viking.

• Tononi, G. et al. (2016). “Integrated information theory: from consciousness to its physical substrate.” Nat. Rev. Neurosci., 17, 450–461.

• Graziano, M. S. A. (2013). Consciousness and the Social Brain. Oxford UP.

• Hafner, D. et al. (2019). “Dream to control: learning behaviors by latent imagination.” ICLR.

• Seth, A. K. (2021). Being You: A New Science of Consciousness. Faber & Faber.

Share:XRedditLinkedIn