Predicate coding: A Model of Neural Encoding and Decoding of Stimuli from Applied Fuzzy-Logic

This work presents a mathematical model of neural coding, according to which neurons encode fuzzy-logic logical predicates applied to stimuli. Moreover, meta-level equations are introduced, which incorporate individual neural codes into a singular unified picture, both for single-neuron codes (such as rate coding and latency coding) and neural network codes (such as population coding and correlation coding). Predicate coding thus parallels the "feature detector neuron" conception, but strengthens it as logical predicates can be made from any well-formed formula containing a free variable (and can thus even model a disjunction of multiple other predicates). This model yields testable hypotheses such as: (i) A neuron's activity predicts the truth-value of a given (fuzzy-logic) predicate applied to a stimulus ; (ii) Each dendrite applies a logical operation over synaptic inputs, producing fuzzy-logic (graded) truth-values for logical connectives ; (iii) Artificially generating code patterns in sensory neurons yield corresponding judgments without stimulus; and (iv) Inhibiting predicate-encoding neurons within a region of the cortex yields corresponding judgments of lack of features, while being presented a stimulus. Hence, predicate-coding, as well as unified equations, provide a mathematically rigorous and experimentally implementable unification of neural encoding and decoding from the cellular scale to large-scale neural networks.