Placebo Response is Driven by UCS Revaluation: Evidence, Neurophysiological Consequences and a Quantitative Model

Abstract

Despite growing scientific interest in the placebo effect and increasing understanding of neurobiological mechanisms, theoretical modeling of the placebo response remains poorly developed. The most extensively accepted theories are expectation and conditioning, involving both conscious and unconscious information processing. However, it is not completely understood how these mechanisms can shape the placebo response. We focus here on neural processes which can account for key properties of the response to substance intake. It is shown that placebo response can be conceptualized as a reaction of a distributed neural system within the central nervous system. Such a reaction represents an integrated component of the response to open substance administration (or to substance intake) and is updated through "unconditioned stimulus (UCS) revaluation learning". The analysis leads to a theorem, which proves the existence of two distinct quantities coded within the brain, these are the expected or prediction outcome and the reactive response. We show that the reactive response is updated automatically by implicit revaluation learning, while the expected outcome can also be modulated through conscious information processing. Conceptualizing the response to substance intake in terms of UCS revaluation learning leads to the theoretical formulation of a potential neuropharmacological treatment for increasing unlimitedly the effectiveness of a given drug.

Figure 1. 1A: responses (understood as plasma GH concentrations) over time (within each trial) and over successive 5 − HT_1B/1D administration trials are simulated.

More specifically, in the first 5 trials the drug is actively administered, so that the implicit learning (and reactive mimicking) takes place, after that, starting from the 6-th trial a placebo is administered. In Fig. 1B the time course of the prediction error over successive trials is computed. Finally, in Fig. 1C the experimental measures and the computational results are compared. In particular, four different curves are reported: a) the overall response after n trials for two consecutive days (experimental6667), b) the overall response after 5 active administration trials (such that the asymptotic response has been reached) from numerical simulations, c) the first placebo response after n active administration trials (experimental measure) and d) the computed placebo response after the 5 active trials. Model parameters adopted for the simulation: T₀ = 150min; τ = 10min; α = 0.45; ; ; . The experimental measured values have been extrapolated from published data (for instance see Figure 5D, pag. 432066).