Is Information in the Brain Represented in Continuous or Discrete Form?
share
The question of continuous-versus-discrete information representation in the brain is a fundamental yet unresolved physiological question. Historically, most analyses assume a continuous representation without considering the alternative possibility of a discrete representation. Our work explores the plausibility of both representations, and answers the question from a communications engineering perspective. Drawing on the well-established Shannon's communications theory, we posit that information in the brain is represented in a discrete form. Using a computer simulation, we show that information cannot be communicated reliably between neurons using a continuous representation, due to the presence of noise; neural information has to be in a discrete form. In addition, we designed 3 (human) behavioral experiments on probability estimation and analyzed the data using a novel discrete (quantized) model of probability. Under a discrete model of probability, two distinct probabilities (say, 0.57 and 0.58) are treated indifferently. We found that data from all participants were better fit to discrete models than continuous ones. Furthermore, we re-analyzed the data from a published (human) behavioral study on intertemporal choice using a novel discrete (quantized) model of intertemporal choice. Under such a model, two distinct time delays (say, 16 days and 17 days) are treated indifferently. We found corroborating results, showing that data from all participants were better fit to discrete models than continuous ones. In summary, all results reported here support our discrete hypothesis of information representation in the brain, which signifies a major demarcation from the current understanding of the brain's physiology.
READ FULL TEXT
