Abstract
Consciousness remains poorly understood as a causative force: existing theories treat it as an epiphenomenal correlate of neural activity rather than explaining how inner experience controls its substrate. I present Recurrent Integration Fractal Theory (RIFT), proposing that consciousness arises when fractal compression of sensory information generates a holographic endospace: the spatiotemporal dimension in which the Self perceives the outer world (exospace) and, through autopoietic feedback, controls the molecular substrate from which it emerged, making consciousness causally efficacious rather than epiphenomenal. In RIFT networks, core neurons form recurrent loops through fractal dendritic trees, generating dynamic information integration through coincidence-based synaptic selection. Coincident EPSPs program somatic multifractals, Ising lattices of ion channels and membrane lipids, encoding a fractal Self-attractor: a geometric field whose coherent point sources generate the holographic endospace through which the Self arises. The Self modulates multifractal growth through lipid domains, controlling ion channel opening probability and action potential generation. Through Generational Fractal Mapping, compressed seeds of prior conscious moments integrate with new EPSPs, replacing infinite downscaling as in classical fractals with sequential self-referential mapping that sustains incremental updating of inner experience, temporal continuity of the Self and Self-attractor transfer across brain regions for global conscious access, establishing irreducibility and unity: the whole is in each part. This architecture was validated computationally against three core properties of consciousness: irreducibility, information integration, and holographic encoding. RIFT generates testable predictions for lipid substrate disruption in Alzheimer's disease, fractal signatures of conscious states, and criteria for consciousness in artificial systems with autopoietic feedback.