## Problem 112 — rank:80

**Title:** Is predictive coding a quantitatively correct brain theory?

**Added binding assumptions (LaTeX):**

```latex
Fix the sensory variable $z_0=y$, the range of the sum, the top-level prior and the neural observation map. The $\gamma_i$ are kinetic gains, typically inverse-time factors after nondimensionalization, not time constants themselves. Gradient descent on a chosen objective is a model hypothesis; matched-complexity predictive comparisons do not prove a unique universal theory of cortex.
```

**Direct edits:** 1. **Fresh external checks:** None; supplied sources retained as provenance.