# A multiple attribute model resolves a conflict between additive and multiplicative models of incentive salience

## TL;DR

This paper proposes a unified multiple attribute model of incentive salience that combines additive and multiplicative effects, simplifying previous dual-structure models and better explaining how stimulus evaluation shifts with internal states.

## Contribution

It introduces a multiple attribute model that unifies additive and multiplicative incentive salience functions, removing the need for dual-structure approaches.

## Key findings

- A multiplicative function suffices for incentive salience when considering multiple stimulus features and interoceptive signals.
- The dual-structure approach with separate weighting of appetitive and aversive features is unnecessary.
- The model explains shifts in stimulus evaluation, such as salt preference changes, without complex function variations.

## Abstract

A model of incentive salience as a function of stimulus value and interoceptive state has been previously proposed. In that model, the function differs depending on whether the stimulus is appetitive or aversive; it is multiplicative for appetitive stimuli and additive for aversive stimuli. The authors argued it was necessary to capture data on how extreme changes in salt appetite could move evaluation of an extreme salt solution from negative to positive. We demonstrate that arbitrarily varying this function is unnecessary, and that a multiplicative function is sufficient if one assumes the incentive salience function for an incentive (such as salt) is comprised of multiple stimulus features and multiple interoceptive signals. We show that it is also unnecessary considering the dual-structure approach-aversive nature of the reward system, which results in separate weighting of appetitive and aversive stimulus features.