Neurons in the brain transmit information to one another through large deflections in their membrane potentials known as spikes. A fundamental question in systems neuroscience is the neural decoding problem: the goal of which is to estimate a sensory stimulus given only the associated neural spiking activity, a task made difficult since such activity is neither independent nor deterministic. As thermodynamic formalism (TDF) is the subfield of ergodic theory concerned with the emergent properties of interacting stochastic elements, one might assume that TDF would assist in determining what impact the structure of interactions would have on decoding. However, the computational intractability of TDF techniques have precluded their application. This talk will detail how a tractable, recently defined approximation of topological pressure can be used to define a non-linear classifier, thereby providing a decoding framework in which the structure of neuronal interactions can be assessed.