Art as Neuroplastogens

Algorithmically generated art can open the same brain-plasticity windows as psychedelics — without the drugs.

The core idea is simple but non-obvious. Psychedelics like psilocybin work therapeutically not because of any specific molecule-receptor magic, but because they force the brain into a computational regime of heightened surprise: the brain's confident top-down predictions get loosened, prediction errors flood upward, and the resulting signal drives rapid rewiring of neural circuits. This is the REBUS model. The paper's claim is that you can get to the same computational regime through the sensory channel — by showing someone the right kind of visual art.

The theoretical backbone is Kolmogorov Theory (KT), BCOM's framework for modeling minds as compression engines. In KT, the brain's Modeling Engine continuously predicts incoming sensory data; when predictions fail, the mismatch (prediction error) does two things simultaneously: it triggers synaptic plasticity (model updating) and, when the error is eventually resolved into a better model, it generates positive affect. Discovery feels good because successful compression is what the objective function rewards. This prediction-error/plasticity/valence triad is the hinge of the whole argument — it's what makes art and psychedelics mechanistically comparable.

Not all art qualifies. The paper formalizes a "Goldilocks zone": stimuli must be surprising enough to keep prediction errors elevated, but structured enough that the brain can eventually compress them. Too predictable and nothing changes; too chaotic and the system gives up and registers aversion. Algorithmic art — fractals, reaction-diffusion textures, latent-space interpolations — is specifically engineered to live in this zone. It preserves the 1/f spectral statistics of natural scenes (inherently compressible) while continuously morphing in ways that violate the current model. The paper also notes that the closed-loop version of this, where real-time EEG tracks neural complexity and valence to keep each person in their personal Goldilocks zone, is something pharmacology simply cannot offer.

The translational target is adolescent depression, a population for whom psychedelics are off the table. The proposed ENAKD/Tx platform pairs EEG-adaptive algorithmic art with cognitive behavioral therapy. The logic mirrors psychedelic-assisted therapy: the art opens a plasticity window, CBT steers the restructuring of maladaptive thought patterns during that window. The paper is honest about what remains unproven — whether sensory-driven prediction errors can match the magnitude of pharmacologically induced ones, and whether daily 15–30 minute sessions accumulate into effects comparable to a single psychedelic dose. These are the empirical questions the platform is designed to answer.